EXERCISE DESERT STRIKE: CONCEPT AND OPERATIONS .. . INCENTIVE CONTRACTING...COMPUTER SCIENCES RESEARCH

NOVEMBER-DECEMBER 1964

THE PROFESSIONAL JOURNAL OF THE UNITED STATES AIR FORCE

E xer c ise Deser t Strike: C oncept and Oper a t io n s ...... An Air University Review Staff Report

Deser t Warfare: A Photographic Portfolio...... 19

Incentive Contracting...... 31 Col. Jack W. Coleman, USAF Maj. David C. Dellinger, USAF

E conomic Aspect s of Mil it a r y Assista nce...... 42 Capt. Douglas N. Jones, USAF

The Science Frontier Resea r c h in Computer Sciences...... 47 Dr. Hans H. Zsehirnt In My Opinion Are We Missing the Boat in Management Devel o pment ? ...... 67 Maj. Robert L. Able, USAF Air Force Review Postgraduate Officer Training for Pil o t Tr a inees...... 74 Capt. Alan J. Grill. USAF

Air Force Officer Education...... 84 Col. John P. Lisack, USAF

Last of the Texa s Towers...... 92

Books and Ideas Tunner and the Saga of Air l if t ...... 95 Col. Raymond L. Towne, USAF (Ret.)

The Contributors...... 99

v-. Address manuscripts to the Editor, Air Uni- the cover versity Review, Aerospace Studies Institute, Maxwell Air Force Base, Ala. Printed by the The vastness of the Mojave Desert became a real Government Printing Office, Washington, and sometimes ominous presence to the more D.C. Subscriptions are sold by the Air Uni- than 100,000 ground troops and airmen fore- versity Book Department, Maxwell Air Force Base, Ala.: yearly S4.50, back issues 75 gathered there in May this year to participate cents. USAF recurrinc pu bi. ic a t io .v 50-2. in Exercise Desert Strike. Air University Review presents a word and picture report on this Vol. XVI No. 1 November-D ec ember 1964 largest of all U.S. Strike Command exercises. EXERCISE DESERT STRIKE Concept and Operations

An A ir University Review Staff Repo r t ®

Cir Uni v er s i t y L i b r a r y . .aswell Air 7orce Base, Ala. PROPERTY U. S. AIR FORCE HE LARGEST United States military United States. As in past joint exercises, tactical training exercise since World War II nuclear weapons training was conducted dur- Twas conducted in May 1964 by the U.S. ing the play of the exercise. Strike Command in adjoining areas of Cali- Appropriately called Desert Strike, the air fornia, Arizona, and Nevada. It was also the maneuver area covered three quarters of a first armored exercise held in the United States million square miles, with aircraft operating since World War II. Two joint task forces with from 28 widely dispersed air bases extending a total of over 100,000 personnel of the U.S. Air from the State of Washington through Texas. Force and Army, over 900 aircraft, and more Incidentally this desert was also the setting for than 500 tanks battled for nearly two weeks on the last training exercises for armored units of a ground maneuver area of some 13 million such mammoth proportions—in 1942—43—when acres in the desert region of southwestern it became known as “General Patton’s Train- ing Center” during extensive ground maneu- •The Editor wishes to acknowledge his indebtedness to Lieutenant Colonel Edward B. Vogel, USA, Directorate of vers there. The Desert Strike ground maneuver Operations ( J—3). Hq USSTRICOM, who assisted in the prepa- ration of the report. area extended from Barstow, California, just 4 AIR UNIVERSITY REVIEW east of Edwards Air Force Base, eastward 170 operating bases. As a result of this emphasis miles to Kingman, Arizona, and from a point and guidance, the air bases acquired for exer- approximately 40 miles south of Las Vegas, cise play included not only Tactical Air Com- Nevada, southward 160 miles to Blythe, Cali- mand bases but other Air Force bases, Navy fornia. This area, consisting primarily of por- and Marine air bases, and civilian airfields. tions of the Mojave Desert and the irrigated Lieutenant General Charles G. Dodge, lands along the Colorado River, was typical of Commanding General, Fifth U.S. Army, com- desert terrain, with extreme differences in ele- manded Joint Task Force Phoenix. His deputy vation, varying from large Hat expanses of was Air Force Major General Clyde Box, Di- sandy desert floor to jagged, rough mountain rector of Plans, usst r icom. jt f Phoenix de- ranges. fended the mythical country of Nezona, gen- Desert Strike was a semicontrolled exer- erally east of the Colorado River, beginning at cise under the direction of U.S. Strike Com- a point southeast of Salome, Arizona, extending mand that allowed opposing joint task forces, north of Kingman, Arizona, through the south- comprised primarily of armored and mecha- ern tip of Nevada, and back to the Colorado nized forces with full air support but including River. airborne units, a maximum of “free play" ini- Lieutenant General Charles B. Westover, tiative to develop, perfect, and test combat Vice Commander, Tactical Air Command, techniques and tactics. Command control was commanded the opposing Joint Task Force exercised only when necessary to ensure meet- Mojave of the mythical country of Calonia. His ing objectives of the maneuver. “Training in deputy was Army Major General Charles H. modern warfare for our combat-ready Army Chase, Special Assistant to the Commander in and Air Force units requires the use of land Chief, usst r icom. jt f Mojave conducted land areas many times larger than areas of even the operations in the Mojave Desert area west of largest military reservations,” said General Paul the Colorado River, the northern boundary D. Adams, Commander in Chief, usst r icom, running from the river west to Fort Irwin, and Exercise Director. The maneuver area California, southeast to Blythe, California, then was selected primarily because the desert ter- back to the river. rain is suitable for large-scale tank movement M ajor General John C. Meyer, Com- and because of its relatively sparse popula- mander, Twelfth Air Force ( tag), at Waco, tion. The dispersal of Air Force units over dis- Texas, was designated as Chief Controller tances similar to those expected in actual (Um pire). He formed his umpire planning combat, together with the freedom of action group at Luke afb, Arizona, on 24 February. which was given the jt f commanders in This group had the mission to determine ex- their employment of ground and air units, actly how Joint Exercise Desert Strike would ensured a realistic no-set-pattern course of be umpired, the command and control proce- combat maneuvers typical of usst r icom exer- dures to be used in umpiring the exercise, and cises. Emphasis was placed on allowing Air to prepare and publish an umpire handbook. Force commanders the maximum amount of The Director Headquarters and Head- flexibility in selecting bases for their fighter, quarters Neutral Forces were established at reconnaissance, and troop-carrier squadrons. Needles, California. This location permitted The initial planning objective was to obtain both these headquarters to be centrally lo- 1/2 air bases per fighter squadron involved in cated within the ground maneuver area and the exercise, which would allow the Air Force was adjacent to a town with adequate accom- commander the flexibility of moving squadrons modations for the observers expected during from one base to another or dispersing squad- the exercise. rons or parts of a squadron among different All commanders were briefed by General bases. This flexibility also permitted the Air Adams and his staff over a two-day period on Force commander to use close-in air bases, the background and purpose of Exercise Des- near the ground maneuver area, as forward ert Strike, the command and control to be EXERCISE DESERT STRIKE 5 used in the exercise, and all the necessary Headquarters, Twelfth Air Force background material to permit them to form 6 fighter squadrons their staffs and develop their initial plans. 1 composite reconnaissance squadron The General Plan for Desert Strike, pub- 1 troop-carrier squadron (C-130) lished on 25 March, was a directive providing 18 Air Defense Command aircraft necessary information for deploying forces to 10 refueling aircraft (Strategic Air the exercise area and outlining the rules of the Command). exercise and plans for redeployment of forces. This was not a tactical operations plan. It was In addition to the two opposing Joint Task very general in providing commanders and Forces, a third force called the Director Re- supporting forces with a background political/ serve was also established. This Director Re- military situation to set the stage for the exer- serve was to simulate a strategic or a theater cise but included no tactical information. Each reserve that might be moved from continental commander also received his operational mis- United States or from strategic staging bases sions from the Exercise Director. These were to the forward combat area. It was planned contained in separate contingency-type plans that all the Director Reserve forces would be presented to each commander, which provided initially located within the Mojave territory him with specific military and political infor- and that on a scheduled basis these units mation and details, including his task organi- would be attached to Joint Task Force Mojave. zation. overall tactical mission, and the con- The Director Reserve initially consisted of the cept of operations. following units: 101st Airborne Division The command structure of Exercise Desert 2d Battalion, 34th Armored Strike consisted of three principal forces: Joint Division Task Force Mojave, Joint Task Force Phoenix, 2d Brigade, 40th Armored Division and the Director Reserve, jt f Mojave initially (California National Guard) consisted of the following units: 6 fighter squadrons 6 troop-carrier squadrons. Headquarters, XVIII Airborne Corps 1st Armored Division For the first time in a large-scale usst r icom 5th Logistical Command joint exercise, major reserve component Army and Air Force units were involved. Air Force Headquarters, Ninth Air Force reserve component units included three pro- 4 fighter squadrons visional fighter squadrons and major elements 1 composite reconnaissance squadron of the 152d and 157th Aircraft Control and 1 troop-carrier squadron (C-130) Warning Wings. Army reserve component 18 Air Defense Command aircraft units consisted of three brigade-size units and 10 refueling aircraft (Strategic Air a number of combat and combat support bat- Command). talions. The 101st Airborne Division, a part of the jtf Phoenix consisted of the following units: Director Reserve, was initially deployed by the Military Air Transport Service to intermediate Headquarters, III Corps staging bases at Mojave-Kern Airfield, Cali- 2d Armored Division fornia, a civilian airfield approximately 20 5th Infantry Division (Mechanized) miles west of Edwards afb, California, and to 258th Infantry Brigade (Arizona) Point Mugu and El Toro Naval Air Stations, (Army National Guard) California. 191st Infantry Brigade (Arizona, By 20 April the initial elements of the Montana, Utah) (Army Reserve) player units moved into the ground maneuver 1st Logistical Command area. They consisted primarily of advance par- \

©Mojave Edwards © A.F5

LosAtigele^,

&an Diego ___ — ©Yima ------^ x \ c. o

In May of 1964, a 22,000-square-mile slash of southwestern desert (light area) was the battleground for U.S. Strike Command’s Joint Exercise Desert Strike.

ties of administrative and logistical support The political situation that gave the exer- units required to meet loaded trains and ve- cise realism and brought about the "war was hicle convoys. Air Force radar installations and a dispute between the mythical countries of ground support units commenced moving to Nezona and Calonia over Colorado River the ground maneuver area and to the air ma- water rights. Also Calonias State of Mojave neuver bases early in May. The Director on the west of the Colorado River had be- Headquarters became operational at Needles longed to Nezona before World War I, and on 9 May. Nezona felt justified in attempting to reoccupy General Nathan F. Twining, USAF Ret., in his role as Prime Minister of Calonia, announces a state of war between his country and Nezona.

Mojave to ensure her control of the river. The dicated by the selection of personnel as par- Organization of Western States had attempted, ticipants in this phase of the play. General unsuccessfully, for ten years to settle the dis- Nathan F. Twining, usaf Retired, former pute between Xezona and Calonia. The exer- Chairman, Joint Chiefs of Staff, was the Prime cise began shortly after Xezona threatened to Minister of the mythical country of Calonia. seize the Parker and Davis Dams, on the Colo- General Jacob L. Devers, usa Retired, former rado River, and the eastern part of the State Commander, Army Field Forces, was the of Mojave. The seriousness displayed by the Prime Minister of Nezona. The war cabinet planners and participants of the exercise and of each of these political leaders was made up the efforts to make the background a realistic of retired and former high-ranking military political-economic-military situation were in- and political leaders. Each held war cabinet CINCSTRIKE T I Director Headquarters

Calonia XXX Joint Control Center XXX-► Nezona War Cabinet Calonia Team/Nezona Team «#-••• War Cabinet ------»------r~ T" I I r - L i 1 ------1______t Director COMJTF Mojave COMJTF Phoenix Reserve

____ » ____ i ARFOR AFFOR ARFOR AFFOR Mojave Mojave Phoenix Phoenix

operational command XXXXX recommendations • •• • d e c is io n s and political/m ilitary policy

The War Cabinets of Calonia and Nezona were each the equivalent of a National Security Council. The Joint Control Center was the equivalent of a Joint Chiefs of Staff organization. The government of Calonia had General Nathan F. Twining, USAF Retired. as Prime Minister, and General Clyde Davis Eddleman, USA Retired, as Defense Min- ister. The government of Nezona had General Jacob L. Devers, USA Retired, as Prime Minister, and Lieutenant General Ira C. Eaker, USAF Retired, as Defense Minister.

meetings twice daily to go over the war situa- ment composed primarily of National Progres- tion, issue political statements, and make sug- sives, the opposition party, who had been agi- gestions. tating to regain the State of Mojave. At 1400 The Joint Control Center (jee) at Direc- hours 15 May diplomatic relations were sev- tor Headquarters had integrated Mojave and ered between the countries of Nezona and Phoenix Operations and Intelligence Teams, Calonia. Late on the same afternoon the Or- which paralleled the responsibility of our own ganization of Western States indicated that it Joint Chiefs of Staff. Tactical situations requir- was withdrawing from the demilitarized zone. ing national-level decisions were presented to As a result, early on 16 May Joint Task Force the appropriate War Cabinet. In turn, direc- Mojave was ordered to seize, secure, and simu- tion and decisions issued by the War Cabinets late operation of the Parker and Davis Dams. were implemented by the je e organization The dams were seized by company-size units through operations orders and frag orders to at 0500 of the same day. Tactical air forces the joint task forces. were restricted from violating the boundaries Both these countries were depicted as with either combat or reconnaissance aircraft having a democratic form of government with prior to the actual initiation of hostilities. a president, prime minister, foreign minister, Calonia’s action in seizing the Parker and and minister of defense. Nezona’s government, Davis Dams sparked offensive action by jt f however, was portrayed as being very unsta- Phoenix, the opposing military force. On 16 ble, with the Conservative Party initially in May Nezona declared war against Calonia and power. On 12 May the president of Nezona issued orders to jt f Phoenix to conduct offen- announced the formation of a new govern- sive actions to seize the State of Mojave. EXERCISE DESERT STRIKE 9

The Desert Strike Cam paign airfields which were so successful that on the morning of 18 May (D + l) Mojave had no Major activities got under way in the exer- available aircraft lor close air support or coun- cise on IS May (D + l) as jt f Phoenix’s 2d terair missions. As a result, by 1700 hours on Armored Division ( “Hell on Wheels ) of D + l (20 hours after the initial attack) jt f World War II fame began to move across the Phoenix had completed three pontoon bridges Colorado River below Parker Dam. The non- and w'as moving the bulk of its heavy armor nuclear attack with all brigades on line and and support units over the river. The crossing onlv battalion-size reserves began in the earl)' was completed successfully through tin- efforts evening of the 17th so as to avoid visual detec- of the 17th Engineers of the 2d Armored Divi- tion by reconnaissance Eights of jt f Mojave sion from Fort Hood, Texas, in setting up a (Map 1). pontoon bridge. Through the first night some 964 pieces of assorted battle equipment moved initial offensive in a steady stream across this structure. Only once, at 0830 the next morning, was the cross- J tf Mojave, faced with two-to-one odds, ing interrupted, and then only briefly when deployed its units in a mobile defense, with two low-flying F-84F’s of jt f Mojave attacked the intention of maintaining dispersion but at and the umpires on the scene declared the the same time being prepared to mass quickly bridge out of action for two hours. to eliminate bridgeheads. The Phoenix tactics The pontoon bridge was a series of rubber of crossing at many points made this plan of boats supporting corrugated-steel planking an- defense difficult to implement, in that the de- chored securely to both banks of the free- fenders had to decide quickly which crossing flowing river. One of the tricks in this particu- to block or delay and which to counterattack. lar crossing was to make sure that all activities Phoenix made maximum-effort attacks with occurred at a time when the gates of the mas- simulated nuclear weapons against Mojave sive Parker Dam to the north were closed and

Map 1. Inaccessible crossing sites on a wide front made JTF Phoenix night attack difficult to counter. 10 AIR UNIVERSITY REVIEW the level of the river was constant The rise tion reached on 18 May was in the vicinity of and fall of the Colorado, with its dams, and Sacramento Springs. Crossings were also re- the swiftness of its flow could play havoc with ported at Parker and Quien Sabe Point. Ne- a crossing. zonan forces moved as far west as the Big As elements of the 2d Armored Division Maria Mountains. Nezonan air forces struck got across the bridge, they picked up speed as far west as Mather afb, California. Air immediately and fanned out on roads and on strikes included low-level strafing attacks and trails into the desert country in their deter- conventional bombing. Working hand in hand mined drive westward to occupy Calonian with the Army ground forces at each river territory. Their immediate objective was not crossing, Tactical Air Command forward air disclosed, but as they fanned out in the desert controllers ( fa c ) arranged for fighter cover to with armored columns in the direction of Rice fly over the bridgehead to ensure that no and Vidal Junction they encountered only further interruption from attacking aircraft in- sporadic slash-and-run resistance. The Mojave terfered with the crossings. forces fell back in the face of the headlong The operation involved the transport of frontal assault. The extreme caution observed tanks and armored personnel carriers across by the forces of jt f Phoenix was indicated by several hundred meters of water by engineer one of the officers of 2d Armored Division’s units in support of the assault. The operation G-3 staff: “Probably the enemy will pick their was made particularly difficult by steep, deeply place to stand and fight. We have to be care- eroded banks and loose soil on both shores. ful to keep our lines of communication open The 258th Infantry Brigade’s participation and not get taken by surprise. W e are contin- in the assault as an element of the 5th Infantry uing reconnaissance by both air and ground Division ( M ech) marked the first time a major to ensure that we know where our enemy is combat element of the Army National Guard at all times and to preclude getting encircled had taken part in a field exercise as an element by the Mojave forces. Some of the elements of an active Army combat division. of the 2d Division ranged as far as 35 miles The ground situation through D + 2 re- west, near Rice. Their scouting parties con- mained fairly stable in the north, with the 5th tinued to probe and push the Mojave forces Infantry Division holding a bridgehead ap- in an effort to open gaps in their defenses for proximately four miles in depth. In the south, advances by the main body of the deploying however, the 2d Armored Division had ex- Phoenix forces. panded its bridgehead to a depth of approxi- Strong armored and mechanized forces of mately 25 miles and was preparing for a major Phoenixs 5th Infantry Division (Mechanized) breakout. During the night of D + l—D + 2, from Fort Carson, Colorado, launched a co- jt f Mojave moved its corps reserve into a ordinated attack across the Colorado River at blocking position along the main avenue of Lake Mojave in a surprise move timed with approach in the south. the more southerly 2d Division’s move west- During the first three days interceptors ward, the immediate objective being the des- from Air Defense Command played a major ert metropolis of Needles, California, and the role in supporting the efforts of each force in securing of Davis Dam. Other units identified gaining air superiority, flying on the average in the northern attack were the 191st and 258th of 2/2 sorties per aircraft per day. Infantry Brigades. Infantrymen of the 258th, supported by their own artillery, crossed the ground/air warfare Colorado River starting at midnight Sunday 17 May (D-day). They used assault boats to This spacious western area was chosen for effect a crossing about ten miles north of Nee- this particular maneuver so that Army heavy dles, after all bridges in the area had been armored units would have room to move and knocked out. The brigade crossing took only demonstrate their maneuverability and driving 1 hour and 20 minutes. The deepest penetra- power. Their moment of triumph came on 20 Map 2. On D + 3 Nezonan forces put a squeeze on JTF Mojave.

May (D + 3) as the 2d Armored Division lenge to the Tactical Air Command in several scored a major breakthrough in the area north- ways. First, it gave tac a chance to operate in west of Blythe by executing a turning move- a training exercise with air bases whose dis- ment, penetration, and flanking attack, causing tances and locations in relation to the combat a complete collapse of the Mojave southern areas were commensurate with those that defensive line ( Map 2). Units of its 1st and 2d could very well be found in an actual war. Brigades had overrun the 1st Armored Divi- Second, it employed simulated tactical nuclear sion command post, capturing the division weapons under controlled conditions. And chief of staff, 69 other personnel, and a num- third, it provided an excellent opportunity to ber of vehicles. demonstrate clearly the flexibility of U.S. tac- Units of the two opposing armored divi- tical air power. sions were intermingled, and the status of the The peril of the 1st Armored Division in defending forces in this area was unclear, the Cadiz Valley on 20 May provided a very though their danger of being overrun was evi- good chance to demonstrate this tremendous dent. In the center, units of the 5th Infantry flexibility of air power. The stage was set Division (Mech) and the 258th Infantry Bri- when the U.S. 2d Division made a surprise gade continued to make steady progress west nighttime thrust against the jt f Mojave flank, of Needles. In the extreme northern sector defended by the U.S. 1st Armored Division. Task Force Linvill, of the XVIII Airborne The 2d Division’s armored column penetrated Corps, Fort Bragg, North Carolina, continued 75 miles on the flank and to the rear of the to defend successfully west of Searchlight unsuspecting Calonia defenders. By dawn, against attacking battalions of the 5th Divi- when the 1st Armored Division commander sion. The task force consisted of two cavalry discovered the situation, the threat was clear squadrons, an airborne infantry battalion, and and unmistakable. The enemy tank columns an armored artillery battalion. had to be blunted immediately, or the entire Exercise Desert Strike was a special chal- forces of Calonia would be vulnerable. The 12 AIR UNIVERSITY REVIEW commander of jt f Mojave’s XVIII Airborne well as a surprise to the invading forces. It Corps was in desperate need of maximum tac- was perhaps even more important, though less tical air support immediately. He called on the spectacular, than the rapid massing of tactical Calonian Air Force, and within a short time air power inasmuch as it gave the fighters lim- a squadron of 18 tac fighters was on hand for itless “staying power” over their ground tar- close air support against the invading armored gets. force. Miles above the ground battle in the But this was a simulated war, and um- Cadiz Valley, aerial tankers of the Strategic pires must be present at an attack or crucial Air Command and North American Air De- situation to score the success or failure of the fense Command were circling. Air National tactical situation. They also determine losses, Guard and Tactical Air Command fighters, damage to equipment, etc. The absence of an giving close air support to the Calonian ground umpire can mean failure to shroud a well- forces, pressed the attack to the limit of their planned and rapidly developing or fast-moving fuel reserves. Then, instead of returning to tactical operation. In the Cadiz Valley engage- base, they climbed to the sac tankers, refueled, ment, the fast-moving tanks of the 2d Armored and returned to the attack. In every war and had outsped the umpires, leaving them far training exercise in the past when a fighter air- behind. Another quick decision was necessary. craft began to run low on fuel, the pilot was The corps commander, wanting to leave no forced to break off the fight and return to base. doubt about the final results of his tactical air Often this meant discharging the rest of his strikes against the tanks, called for air rein- ammunition against targets of opportunity on forcements. Within 15 minutes a second squad- the way home or landing with it unused. This ron of 12 fighters, operating from a distant was not the case in Desert Strike. base, was on the scene, giving the Calonia de- More than S million pounds of jet fuel fenders a mass of 37 supersonic tactical fight- were transferred in aerial refuelings during ers to smash 50 advancing tanks of the 2d Desert Strike. At no point were attacking Armored Division. The results were definite— “enemy fighters able to “destroy” enough of the invading force was destroyed and the flank the tankers to halt air refueling operations. of Calonian forces was protected. One F-100 operating from Biggs afb, Texas, The odds were very uneven, of course: 37 refueled five times from a KC-135 and was fighters against 50 tanks. But the situation was airborne for 6 hours 20 minutes. desperate, and time was a factor. The high Another advancement over earlier war- mobility of supersonic aircraft was the means fare techniques used in Desert Strike for the of providing overpowering odds within min- first time was night tactical air warfare. Dur- utes. Then, too, the missing umpires had to be ing World W ar II and the Korean War tactical convinced. General Meyer, Chief Controller air operations were limited largely to the day- (Um pire), explained the situation: "It would light hours. In Desert Strike large-scale tacti- not be possible or desirable to mass an effort cal air operations were carried on by Calonian of this size against each and every ground defenders at night in the application of a new target. VVe must have sufficient land combat tactic, dubbed “Night Owl." This technique forces to deal with most threats of this kind, consists of a continuous attack by flare- but we know that we have the means to deal equipped jet fighters. The flares used are so with a major emergency, such as occurred in far advanced over those of World War II and the Cadiz Valley.’ Korea that they can light an area as large as A new technique, recently developed by an air base. Each fighter element explodes a the Air Force and tested in a tactical situation new flare as it completes its attack, illuminat- for the first time during Desert Strike, helped ing the target for the next element. to give the tactical fighters more punch during The reconnaissance essential to identify this encounter with the 2d Armored Division and assess the importance of such targets was and was an important factor in the victory as gathered, day and night, by tac: RB-66 and The M-60 machine gun (above) mounted in anti- aircraft configuration. . . . E v en b y t h e d a r k o f night swift-moving M-60’s of JT F Phoenix’s 2d Armored Division invaded ]TF Mojave territory. 14 AIR UNIVERSITY REVIEW

RF-101 aircraft using photography and more on 21 May, a major threat in the southern exotic infrared and electronic means. A com- sector was averted on the same day when plex up-to-date system of strobe lighting equip- nuclear strikes by Honest John missiles on the ment was used on two RB-66B aerial recon- 2d Brigade, 2d Armored Division, caused this naissance aircraft flying night photoreconnais- force to withdraw 25 kilometers to organize sance missions out of Norton afb, California, and recover. The sting was definitely felt by for jt f Mojave. The light system is installed the invading forces, and jt f Mojave sensed a in the aircraft bomb bay, and each flash has a change in the tide of battle and threw addi- duration of one-thousandth of a second. It is tional reserves into the conflict and at the set off when the pilot triggers the camera same time began mobilizing additional Na- shutter. More than 35,000 watts of electricity tional Guard and Reserve forces. flow into each light, providing enough illumi- During this period two airborne brigades nation for clear aerial photographs of enemy were airlanded at assault airstrips to rein- ground targets. force the Mojave defensive line. Over 85 C-130 aircraft airlanded 8000 troops and 3500 tons counteroffensive of equipment in the forward combat area in less than 36 hours, using a total of three strips Early on the morning of 21 May (D + 4) (M ap 3 ). The primary assault landing strip Exercise Desert Strike escalated into nuclear used was 15-mile-long Bristol Dry Lake, on warfare as the Calonian Air Force launched which the C-130’s practically selected their the first simulated nuclear attack of the war. own runway. Using a unique system on this At about 0500, Nezonan air bases at Luke, strip, the C-130’s touched down on one side Davis-Monthan, and Yuma in Arizona, Biggs of the lake, made a 90° right turn, offloaded, in Texas, Hill in Utah, and Nellis in Nevada turned left again, and took off. This system per- were struck with nuclear weapons by tactical mitted a touchdown rate of one aircraft every fighters. Succeeding strikes destroyed the two minutes. bridges across the Colorado River at Needles The Nezonan Air Force was capable of and north of Davis Dam. Finally, strikes were striking back, of course. About noon on 21 May directed against the advancing columns of the simulated nuclear weapons hit Calonian air 2d Armored Division in the vicinity of Cadiz bases at McClellan, Mather, Hamilton, Le- and Bagdad. The results of these attacks were moore, Edwards, Palmdale, Oxnard, Norton, decisive: Luke—out of action for 24 hours and and George, all in California. The strikes were severe damage for 8 hours to 2 fighters and 19 not decisive because Nezona’s air capability reconnaissance aircraft; Nellis—extensive dam- had been reduced earlier by the Calonian age to 3 fighters and to hangars, buildings, and forces’ attacks. Both sides were then engaged supplies; Biggs—severe damage to 2 fighters in nuclear war, but Nezona forces were being and extensive damage to vehicles; Davis- pushed back off Calonian soil as a result of the Monthan—base out of action for 24 hours, initial devastating nuclear surprise attack. severe damage to 12 fighters, and extensive On Thursday 21 May the Prime Minister damage to vehicles; H ill—minor damage; of Calonia, General Nathan Twining, issued Yuma—severe damage to fighters and exten- the following explanation for Calonia s ini- sive damage to vehicles, supplies, and petro- tiating the use of nuclear weapons in its attack leum stores. on Nezonan air bases and battlefield targets: The decisive victory of the Calonian Air Nezonan forces in overwhelming strength Force over the 2d Armored Division’s heavy launched a two-pronged attack against Calonia tanks in the Cadiz Valley on 20 May seemed Sunday and throughout Monday. As the world to trigger a chain of events in which the jt f knows, this attack was unprovoked and unwar- Mojave evidently decided to go all out to turn ranted. When Calonia publishes the record of the tide of battle. In addition to nuclear strikes correspondence between the respective war by fighter aircraft on jt f Phoenix’s air bases cabinets, which Calonia terminated on Monday >

Map 3. Nuclear strikes lead the ]TF Mojave counterattack on D-j-4.

after Nezona duplicity was conclusively estab- trative break, jt f Phoenix resumed its attack, lished. the world will appreciate that Nezona’s and by the morning of D + 7 it had success- act of naked aggression must, along with its fully penetrated the Mojave defenses in two perpetrators, be condemned and defeated. To areas. By 1500 hours on D + 7 the Phoenix allow such aggression to succeed, to permit attack stalled because of overextended lines of Nezona to gain and hold one square inch of Calonia territory, would be to repudiate the communication, jt f Mojave seized this oppor- values of the civilized world and the moral tunity to launch a counteroffensive to drive standards of the Organization of Western Phoenix back across the river. States. At 0620 on D + 8, a brigade of the 101st In an effort to stem Nezonan attack and to Airborne Division was parachuted 30 miles mount our own counterattack, we have found behind Phoenix lines to seize a critical pass it necessary as a last resort to utilize nuclear and await the linkup of the armored forces weapons. We informed the Secretary-General (Map 4). A total of 67 C-130 aircraft was em- of the ows on Monday that we would make ployed in this drop, 27 carrying troops, 40 every effort to confine the conflict to the tactical battlefield and to avoid escalation to nuclear carrying heavy drop. This was a very success- weaponry. We held out as long as possible and ful operation, with only one abort and mini- authorized their uses yesterday afternoon. We mum injuries to the paratroopers. A follow-up have employed small weapons against military airdrop ran into difficulty when attacking targets only, have authorized air bursts only in Phoenix fighters struck the staging base while an effort to minimize fallout, and have in- the airborne troops were loading, destroying structed our commanders to take every pre- over 31 aircraft and inflicting heavy casualties caution to protect non-combatants, Calomans on the crews and airborne troops. and Xezonans alike. We are informed that fall- sa c KC-135 tankers, attached to each jt f , out has been minimum and that Calonians can go about their daily tasks without fear of flew maximum effort during this period to per- hazards to their health. mit deep penetration by fighters on counter- air missions. On D + 8 Mojave tankers com- At noon on D + 6, after a 24-hour adminis- pleted over 150 mid-air refuelings in support c)TP MOOAMe ^ ^> v < J r s r Map 4. The counteroffensive con- ? tinues on D + 8 as JTF Mojave pushes ICO Nezonan forces toward the river.

of offensive actions. weapons against the strong points of Nezona, By dusk on D + 8 Mojave had successfully but they were caught off balance by the in- recaptured a bridge over the Colorado River crease in the tempo of attacks launched by and had forced Phoenix into a full withdrawal. jt f Phoenix. Further adding to the new punch In the north the attack by the 101st Airborne of Nezonan forces was the reconstitution of Division drove the opposing 5th Infantry Di- both land and air forces by expanding mobil- vision back across the river into its homeland. ization. In the south, however, the two armored divi- sions continued to clash west of the river. The Mojave Reconnaissance Squadron T he pa r t ic ipa n t s in Joint Exercise Desert (RF-101 and RB-66) was used quite exten- Strike were aware that the planned deadline sively during this period to locate bypassed for the cease fire was drawing near, but they Phoenix units attempting to extricate them- did not know the exact hour. Each jt f com- selves. This was a particularly difficult mis- mander, with his staff, was trying desperately sion, since these forces had to be distinguished to register with the umpires, score points, and from friendly Mojave ground forces. win the exercise. They tried to pull surprises, jt f Mojave continued to push hard against catch the opponent off guard and off balance, Nezonan forces after the initial nuclear strike gain ground by use of nuclear weapons or gave it an edge, and on 26 May (D + 9) the mobile maneuvering or concentration of force. 5th Infantry Division of jt f Phoenix started jt f Phoenix fought to keep the Calonian forces to withdraw across the Colorado River. Also off Nezonan soil. All these efforts brought on that date the Nezonan Air Force launched about major changes in the tactical situation an all-out nuclear attack on air bases of Cal- right up to the cease fire at 1700 on 29 May. onia in addition to five nuclear strikes against At that time almost all elements of the 5th enemy positions on the battlefield. Calonian Infantry Division had recrossed the Colorado forces continued their attacks with nuclear River, reinforcing the northern sector of the - » -»► ^ i

Map 5. JT F Phoenix teas driving hard to regain the offensive as the desert war games terminated on D-rl'2. M )(

pocket held by the 2d Armored Division. The cises such as Desert Strike. There is simply exercise controllers believed that the tactical no adequate substitute for this type of realistic commanders had had ample opportunity for evaluation.” free play and that the objectives of the exer- General Adams, in a statement summa- cise had been achieved to a degree equal to rizing the exercise, said, “It will be a matter or surpassing all expectations. At that point of months before all the data obtained during there was no need to prolong the exercise and the course of Joint Exercise Desert Strike can mount the cost (Map 5 ). Air and ground units be properly evaluated. Complete and sound received outstanding training in joint opera- conclusions based solely on what one or a tions associated with the simulated use of both few men can derive from necessarily limited conventional and tactical-nuclear weapons. observations are not possible at this time. This exercise provided a rare opportunity for There is however no question in my mind evaluating tactical air action in support of a that through a process of careful assimilation highly fluid, fast moving ground situation. and evaluation of data we shall obtain from Tactical air action in support of ground forces this maneuver guidance far beyond anything was extensive. The Mojave ground forces re- we had anticipated a few weeks ago.” In his quested 1636 sorties, of which 1628 were flown praise of the reserve forces General Adams and 1500 judged effective. Phoenix ground added, “I am gratified by the results achieved forces requested 1246 sorties, of which 1163 by reserve component units participating in were flown and 1080 judged effective. the Exercise. We literally have a ‘gold mine’ “Out of this huge Desert Strike effort.” of information which when refined will pro- General Meyer said, “we have advanced the vide us a large body of highly valuable knowl- defense of the United States, while identifying edge for use throughout the services.” many areas in which much greater concentra- Reliable sources have estimated that Ex- tion and study is required. As progress is made ercise Desert Strike cost about $54 million, in those fields, it must be tested in field exer- the costliest exercise since World War II. 18 AIR UNIVERSITY REVIEW

Asked why the exercise was considered worth simulate realistically the employment of tac- the outlay of funds, Pentagon officials cited tical nuclear weapons. This let us better explore these reasons, among others: tactical use of nuclear weapons in a limited war situation. It is the first time since World War II that Joint Exercise Desert Strike was not only a we have been able to get two armored divisions thorough test of men and equipment, but also into an exercise maneuver area in the United an ideal opportunity to test and further improve States where they could operate as freely as several operational concepts for joint air- required for proper armor field training. ground operations we have been developing The exercise also gave us a chance to and improving over the past several years. DESERT WARFARE

A PHOTOGRAPHIC PORTFOLIO Desert Ground forces of Nezona had lo cross a river before they could chal- River Crossing lenge the forces o f Calonia and of INature in the vast Mojave Desert.

A .smoke screen is set off to hide an assembly area for vehicles and sol- diers of JTF Phoenix preparing to cross the Colorado River into Calonia. After crossing the river in rubber rafts by moonlight, National Guardsmen climb the “enemy'’ embankment. . . . \1- 116 armored personnel carriers swim the river and head for the “front,” leaving the Ne- zonan homeland far behind.

Pontoon bridge construction proceeds from both sides toward the middle. . . . Xl-60 main battle tanks of the 2d Ar- mored Division are first across the newly completed floating pontoon bridge. Desert U.S. Army’s iron monsters had plenty of maneuver room in the Mojave Ground Warfare Desert, where sun, sand, and wind were a real threat in the mock war.

The task commander of the Calonia forces checks terrain for the best route through the Mojave. Soldiers and airmen get ready for the day’s work at a camouflaged bed site near JTF Mo- jave headquarters. ... A JTF Mojave vehicle park lies hidden in the foothills of the Old Dad Mountains north of Amboy, California. An M-60 of 2d Armored (“Hell on W heels”) Division, its 105-mm howitzer pointing the way, plows into a sea of desert dust. The 2d led the Nezonan attack on Calonia. ... Al-116 armored personnel carriers of JTF Phoenix move Nezona’s forces into battle.

K Desert Fighter aircraft made it hot for ground forces round the clock as new tech- Air Warfare niques enabled the first large-scale night operations in close air support.

Rows of equipment parachute-packed for airdrop with 101st Airborne Division paratroops await loading onto C-130 transports at Kerns County Airport, Mojave, California. The trailers are backed to the aircraft, and the pallets are pulled over rollers for easy unloading. /

» r

A Calonian F-100 Super Sabre makes a low-altitude simulated \ bomb run against JTF Phoenix 379c; troops along Route 66 near Am- Y boy, California. ... A C-130 lands on impacted runway pre- pared by Army engineers in the bed of Bristol Dry Lake. An Air Defense Command F-101B Voodoo prepares to take off on a Desert Strike interceptor mission.... On the heli- copter pad at Headquarters JTF Mojave, an Air Force H-19 and Army UH-1B meet. An F-84 Thunderstreak zooms upward after a bomb run against Nezonan armored columns. INCENTIVE CONTRACTING

C olonel J ack W. C oleman and M ajor David C. D el l inc er

ODAY, as has been true historically, of research and development and its concen- defense procurement emphasis is on tration on realistic project definition is to con- T competitive procurement. True compe- tribute to the achievement of such a profit ob- tition usually results in a firm-fixed-price con- jective. Success in these efforts will permit the tract under which the parties have in effect use of contracts that will also contribute to the agreed that the contractor has assumed full same overall objective. By the judicious use of responsibility, in the form of profits or losses, contract incentives on cost, performance, and for all costs under or over the firm fixed price. schedule, it is believed that the Government Unfortunately, a large majority of the Depart- can establish the desired emphasis and criteria ment of Defense procurements (particularly under which achievement can be considered research, development, test, and evaluation) outstanding, mediocre, or poor. When profits do not lend themselves to the ideal competi- are tied to contract results, contractor motiva- tive situation. The nature of many of the tion will be maximized. While industrial reac- weapon system procurements results in cost- tion to Governmental efforts to promote ef- type contracts with fixed-fee arrangements ficiency through contractual incentive arrange- (cost plus fixed fee—cpff). In this type of ments has been mixed, one spokesman has contract the contractor has little if any stake stated: “Here again use of fear, inspirational in the outcome. Since the fee is fixed, con- leadership, appeal to pride and patriotism, in tractor cost responsibility is minimal, and in appropriate cases, can have an effect, but effect good performance and bad performance basically the appeal to the pocketbook gets the are equally rewarded. Thus there is little in- mostest for the leastest.”1 centive to control costs or improve perform- As deficiencies have been recognized in ance and schedules. various types of contracts,'- greater emphasis Both the Government and the contractor and interest have been placed on the use should be concerned with having the profit of incentive provisions within such contracts mptive work for the truly effective and eco- to correct these deficiencies and actively en- nomical performance so vital in the interest courage the reduction of costs to the Govern- of national defense. The objective should be to ment. The concept is not new,3 but the em- ensure that the efficient producer is rewarded phasis is new and is a key element in the evo- by high profits, the mediocre producer by lution taking place in procurement policy and moderate profits, and the poor producer by practice. Various management consultant firms low profits or losses. One of the objectives of and Governmental committees, in their analysis the Department of Defense in its recent efforts of this problem of promoting efficient con- to define more clearly the various categories tractor performance, have recommended 32 AIR UNIVERSITY REVIEW greater use of incentive contracts. One such define fpif and cpif contracts and restriction group stated: “It is believed that permitting on fee as follows: contractors to be rewarded pricewise and to share in savings for good performance and The fixed-price incentive contract is a fixed-price type contract with provisions for economy of operation would be an effective adjustment of profit and establishment of the means to improve procurement.”4 The Depart- final contract price by a formula based on the ment of Defense has defined the incentive relationship which final negotiated total cost principle as follows: bears to total target costs. The incentive principle holds, in brief, The cost-plus-incentive-fee contract is a that a contractor should be motivated, in cal- cost-reimbursement type contract with provi- culable terms (i) to turn out a product that sion for a fee which is adjusted by formula in accordance with the relationship which total meets significantly advanced performance allowable costs bear to target cost. . . . The pro- goals, (ii) to improve on the contract schedule vision for increase or decrease in the fee is de- up to and including final delivery, (iii) to sub- signed to provide an incentive for maximum stantially reduce the costs of the work, or (iv) effort on the part of the contractor to manage to complete the project under a weighted com- the contract effectively. bination of some or all of these objectives. ’ 10 U.S.C. 2306 (d) provides that in the Emphasis is on getting the lowest total case of a cost-plus-fixed-fee [emphasis added] price consistent with performance and deliv- contract the fee shall not exceed ten per cent ery objectives rather than on limiting a con- (10?) of the estimated cost of the contract, ex- clusive of the fee, as determined by the Secre- tractor to the lowest possible profit.13 tary concerned at the time of entering into such While there are influential proponents of contract (except that a fee not in excess of the incentive concept, the concept also has in- fifteen percent (15%) of such estimated cost is fluential opponents. Chairman Vinson of the authorized in any such contract for experi- House Armed Services Committee is on record mental, developmental or research work and as opposing the use of incentive contracts be- not in excess of six percent (6%) of the esti- cause the increased fee can result from poor mated cost, exclusive of fees, . . . in contracts negotiation of target costs (which form the for architectural or engineering services re- basis for determining reward or penalty) as lating to any public works or utility projects.) easily as from increased operational efficiency. [ aspr has restricted maximum fees in cpif con- tracts to these same tenets.—aspr 3-405.4( c )] Vice Admiral Rickover, in testimony before As to fee limitations on subcontracts, see aspr the House Defense Appropriations Subcom- 3-807.10(d).8 mittee on 23 May 1963, expressed the sarpe criticism and challenged the widespread use The following elements, by type of con- of incentive contracts: “The contractor knows tract, are subject to Government and con- more about how much the item costs than the tractor negotiation before award: government, and can set a high initial price. The contractor can then say his lower costs Element FPIF CPIF were due to efficiency . . . and the government Target cost X X rewards him for these so-called cost savings by Target profit X x ( fee v: paying him a higher profit.”7 profit) Ceiling price X Sharing formula X X elements of incentive contracts Maximum fee X X Generally, two broad types of contracts Minimum fee form the basic contractual vehicles for incor- porating incentives. These are the fixed-price- A discussion of certain of these elements incentive-fee contract ( fpif) and the cost- is necessary before proceeding. plus-incentive-fee contract ( cpif). Tartlet Cost. As early criticism has indi- Armed Services Procurement Regulations cated, the most critical aspect of incentive INCENTIVE CONTRACTING 33 contracting is to arrive at a realistic target cost. in arriving at these limits are as follows: Unfortunately, since target cost is an estimate, (1) The nature of the work. The higher a forecast of the future, it will inevitably re- the-ratio of development to fabrication, the flect some degree of uncertainty as to the spe- greater the uncertainty. cific outcome of the contract. The target cost (2) Past experience. No previous experi- should represent the best, mutually deter- ence in type of work or poor cost experience mined estimate of what actual costs will be in forecasting will increase the uncertainty. when the contract is complete; or, stated an- (3) Negotiation environment. If critical other way and borrowing from statistics, tar- cost positions of each party are far apart ( and get cost should represent that figure at which not caused by misunderstanding of work there is an equal probability of costs being statement), one (or both) is likely to have either under or over. It goes without saying little confidence in the end result. that these are not easy criteria to meet. How- (4 ) Time available for development of es- ever, with more realistic project definition and timate and negotiation. The more time per- such procedures as program evaluation and mitted, the greater the confidence in the end review technique and PERT/Cost, it is not im- result.10 possible. It should be recognized that all is not The Vinson and Rickover criticisms con- lost even if the criteria cannot be completely cerning target costs noted earlier must assume complied with. The confidence that either or that the Government can and does negotiate a both parties have in target will affect negotia- better position in contracts which exclude in- tion of the other elements. Statistically, the centive provisions. We are not aware of any exact cost estimate, while important, gains analyses or studies that support such a conclu- significance as the range of possible costs nar- sion. Rather, we believe that in many cases rows towards the fulcrum (or target cost in for many reasons the Government cannot ne- this case), because confidence varies inversely gotiate the optimum cost (regardless of inclu- with such range. If the Government has less sion or exclusion of incentive features). How- than the desired confidence in target (e.g., ever, it would still appear a distinct advantage there is a greater chance of overrun than un- to the Government to provide some incentive derrun), it can negotiate stronger incentive to get actual performance to correlate with provisions and the type of contract that will a more efficient operation. For example, if the minimize this aspect. Government has negotiated the best position In arriving at a mutually agreed-to target it can but this is not very good, it would still cost, the negotiations of necessity will also pro- seem prudent to try to reduce total cost by vide data to support each party’s confidence some sharing arrangement. A share arrange- level and the extent of cost range. The lower ment of 75/25 would mean that if the con- limit of this range would reflect the most opti- tractor is motivated to reduce costs below the mistic or lowest practical cost for accomplish- excessively high target, the Government would ment, whereas the upper limit would reflect recoup 75 cents of each dollar of reduction. the most pessimistic practical cost for accom- While maybe not optimal, this is a more frugal plishment. This would in effect establish the position than would have been achieved with- range of confidence, dod has indicated that, out incentives, because there is considerable “When the upper limit is less than 10S, the evidence that without some incentive the con- Government should concentrate on negotia- tractor will, at minimum, incur costs up to the tion of a firm-fixed-price arrangement. For an dollar limit cited in the contract if it is other upper limit between 10 and 25 percent, the than a firm-fixed-price contract. fixed-price incentive type will usually be ap- Target Profit. Target profit (or any nego- propriate. When confidence decreases to a tiated profit for that m atter), like target cost, level beyond 25 percent, it will usually be has proved most difficult to arrive at in any necessary to shift to the c pif form.”'' ( Skewing realistic and satisfactory manner. This area, in of range can be compensated for through the terms of supporting documentation and time incentive formula.) Factors to be considered consumed in negotiations, seems to be taking 34 AIR UNIVERSITY REVIEW more and more effort.11 In many instances, to have a relatively shallow share line, say after extensive negotiation, the resulting profit 85/15 within a small range of target cost (if or fee has, by some coincidence, usually ap- small deviations are unimportant), and steeper proximated the appropriate industry average. slopes for costs farther from target if large over- To make profit or fee more indicative of risk, runs are objectionable and it is desirable to en- Section III of Armed Services Procurement courage substantial underruns. ( However, this Regulation has been revised to incorporate a so-called “plateau” effect should not be a “gim- weighted guidelines concept which became mick” to surmount impasses in the negotiation effective 1 January 1964. Again, this concept of target cost.) is an effort to reward good performance and penalize mediocre or poor performance by having the profit objective reflect the total maximum-minimum fee overall task to be performed and the risk as- Obviously, the objective of maximum- s u m e d b y t h e c o n t r a c t o r . 12 C l o s e l y a s s o c i a t e d minimum fee is to have a reward-penalty ar- with this is the new dod system for evaluating rangement that will provide the contractor contractor performance.13 with a valid incentive to control costs. The in- Sharing Formula. If target cost and ex- centive provisions should be so constructed as pected range and target profit or fee and to be effective over the entire range of possible maximum-minimum fee are properly devel- deviations from target cost. The Department oped, then in effect the sharing arrangement of Defense indicates that the minimum and has been predetermined. This is proper inas- maximum fees and the fee adjustment formula much as the most important determinant of should be negotiated so as to provide an incen- the "share slope should be the degree of risk tive which will be effective over variations in inherent in the particular procurement situa- cost throughout the full range of foreseeable tion. Further, share arrangements and thus variations from the target. Further, dod sug- slope can be tailored to meet specific conditions gests that when high maximum incentive fees and do not have to result in a continuous are negotiated, it is proper (for control at the straight line. For example, it may be desirable other end of the spectrum) to provide for a

• cost (millions)

Figure 1. Fixed-price-incentive-fee (FPIF) contract INCENTIVE CONTRACTING 35 low minimum fee which may be a “zero" fee d. The intersection of the 80/20 and or a “negative” fee.14 0/100 share lines is the total dollar point w'here the contract starts operating as a firm- fixed-price contract. In the example this point fixed-price-incentive-fee contract is at $1 12.5m . ° What would the total contract In fpif contracts, cost reductions or over- price be if costs equaled S90m ? From the runs from the target cost are shared by the graph the fee can be determined by adding it Government and the contractor in accordance to the actual costs to get the total price. The with a negotiated formula. The distinguishing fee for $90m of costs equals $12m .°° Thus feature of this form, however, is that of a fixed total price equals $102m . upper limit on the total contract price, i.e., Note that, as constructed, profit can range cost plus profit. This feature has the effect of from a negative value ( if actual allowable changing the characteristics of the contract to costs exceed ceiling) to a zero profit (if actual that of a fixed-price agreement. With this allowable costs equal ceiling price) to any background it is desirable to portray briefly positive amount on the share line (if allowable the conditions by type of contract. Assuming costs are less than ceiling price). The ceiling the following facts associated with an fpif price, which represents the maximum amount contract, one can graphically portray types of the Government will pay, is a very desirable incentive arrangements. (Graphic portrayal feature and, other things being equal, makes it has distinct advantages in that it permits preferable to a cpif type. quicker and greater understanding of the arrangement and any desired manipulation thereto. See Figure 1.) cost-plus-incentive-fee contract .Assume: Next, with the following assumed facts, it a cpif Target cost SI 00m is possible to portray contractual ar- rangement graphically'. Target profit SIOm Assume: Ceiling price SI 20m Sharing formula 80/20 Target cost SI 00m Target fee S 10m To construct the graph: Maximum fee S 13m a. Plot target cost against target profit. Minimum fee S 7m b. Reduce (or increase) target cost by Fee swing ±$3m a certain number of dollars (e.g., SIOm ) and Sharing formula 80/20 calculate the profit (e.g., by reducing cost The same mechanics applicable to construct- m m SIO , i.e., to $90 ), profit would equal $10.\i ing the fpif contract can be used here except 4- (202 X SIOm ) = S12m . Plot S90m cost vs. that no ceiling price is involved and, instead, $12m profit to establish second point for there are maximum and minimum fees. Under straight-line share arrangement. actual conditions the confidence level devel- c. At some point, the share line changes oped through analysis and negotiation would from 80/20 to 0/100. To find that point graph- ically: °This may be computed mathematically as follows: (1) Final point equals ceiling price, i.e., ceiling price — contractor’s % share Breaking Point X = " U W ?Pst ~ P™fit profit equals zero at cost equal to S120m . 1 — contractor s % share (2) The second point is computed similarly substitute: = ~ ’2 <»«” «> ~ to b, above, except that on a 0/100 share line 1 - .2 any cost reduction equals a corresponding in- = $90 Si crease in profit to contractor. Assume reduc- .8 tion of costs from S120m to SIIOm ; then profit = S112.5S would be SIOm . P l o t $ 1 0 m vs. SIIOm and •“This can be calculated based on sharing arrangement: •S90m actual costs is $10Sr less than target; therefore, the con- draw straight line. tractor fee will be $10si + (20% X $ 10m ) = $12Si. Figure 2. Cost-plus-incentive-fee (CPIF) contract

produce a range ( most optimistic, most pessi- portrayed graphically; e.g., assume that at 10% mistic, target cost), and extending upward plus-minus target cost it is desirable to have a from these costs to an intersection of the 90/10 share arrangement with cost less than "max-min” fee lines should be the technique that sharing 75/25 and cost beyond that shar- for actually determining the share formula. ing 80/20. Further assume the same additional ( If most optimistic and most pessimistic costs facts used above except that minimum fee is approximated $85m and $115m , the share line shifted to $5m . Figure 3 graphically portrays would be as given above, i.e., 80/20.) What this situation. This contractual arrangement would be the total price of the contract to suggests that slight (10%) variations from tar- the Government if contractor costs equaled get are acceptable, but the stronger incentive $112.5m ? From the graph (Figure 2) you can is to get costs below S90xi and a fairly good find the fee at that cost, which is $7.5m ,° penalty if they exceed $110m up to $130m , giving a total contract price of $120m . at which time the cost control incentive ceases. Note the major difference between the two contractual arrangements. The cpif with multiple incentives a minimum fee becomes a cost-plus-fixed-fee contract when costs reach the point where In addition to the cost incentives already minimum fee becomes applicable ( i.e., if costs discussed, there is a growing emphasis on the overrun target to point of minimum fee, no need to provide an incentive to the contractor further penalties are imposed and the control in the other major areas over which he has incentive feature is nullified). As noted earlier, control, i.e., performance and schedule. These the fpif contract reverts to a firm-fixed-price three broad areas—cost, performance, and ( ffp) contract, thus providing maximum con- schedule—are involved in any satisfactory con- trol over incentive. tract completion, and emphasis on one to the If it is desirable to have other than a exclusion of the others risks loss of control straight line share arrangement, it is easily regarding the desired objective of the Govern- ment. After all, the timely delivery of a weap- “The fee can be calculated based on the sharing arrange- ment. E.g., if costs exceeded target by $12.5m , this would cost on system evidencing required technical capa- the contractor $2.5Si of fee (20% X •$ 12.5m ), making fee $7.5 Sr. bilities is no less an objective than cost. The cost imillions

Figure 3. CPIF contract with cost /share variations

Department of Defense has indicated that a Historically, most schedule incentives properly structured multiple-incentive con- have been negative in nature. That is, no re- tract should serve two basic purposes: ward for early delivery but penalties for fail- ure to perform on time. However, if early deliv- First, it should motivate the contractor to strive for outstanding results in all three incen- ery of finished work has a value to the Gov- tive areas; . . . Second, if it becomes apparent ernment, there is no reason why a positive to the contractor that outstanding results can- incentive arrangement should not be in- not be achieved in all areas, the incentive cluded.17 The success of schedule incentives, structure should compel decisions as between as is true with all incentives, is largely de- costs, time and performance that are in con- pendent upon the development of realistic tar- sonance with the overall procurement objec- gets. Secondly, the range of incentive effective- tives of the Government.15 ness must be established18 (i.e., the time Performance incentives (reward/penalty) period preceding and following these targets normally are associated with such things as during which the incentive will be operative). weight, speed, reliability, range, payload, etc. At best, structuring a multiple-incentive Assuming good project definition,10 the tech- matrix so that the most profitable trade-off de- nical aspects of the equipment and procure- cision for the contractor will always coincide ment should be clear. This will form the basis with the desired objectives of the Government for determining those elements of performance is most difficult. Technically, the relationship where emphasis is desirable, defining what of these elements ( cost, performance, and constitutes superior technical achievement, a schedule) is much more complex than is recog- reasonable target, exactly how performance nized by most, and there is considerable ques- will be measured, and the relative importance tion as to whether these relationships can in of each characteristic. Like cost targets, per- fact be related in some effective way. Conse- formance targets must be reasonable, and if quently, extreme caution must be exercised in trade-offs become necessary between perform- developing a multiple contractual arrange- ance elements, the incentives should be struc- ment, or more harm than good may possibly tured so that contractor emphasis is in con- be done. sonance with Government objectives. As an example of multiple incentives and 38 AIR UNIVERSITY REVIEW

their graphic portrayal, assume a CPrF con- tate a more precise statement of the con- tract and the following facts: tractor’s position regarding this particular ex-

Target cost $100m ample, the following notations and calcula- Target delivery 24 months after tions are provided. Let: contract award X = Miles per hour that actual perform- Performance ance exceeds 950 mph (minimum ac- target-speed 1000 mph19 ceptable performance) Target fee $7m (77) Y = Millions of dollars that actual cost ex- Target fee swing ±$8m ceeds $70 million (estimated best The ranges of incentive effectiveness have cost possible) been determined to be as follows: Z = Number of months that actual sched- ule exceeds 21 months ( estimated Incentive Required for Required for shortest obtainable schedule) Element Max. Reward ur"( f Max. Penalty Fr= T otal fee ( target fee dt incentive). Cost $70m 8100m $ 130m Performance 1050 mph 1000 mph 950 mph Schedule 21 months 24 months 27 months Further, the Government believes that per- Figure 4. CPIF contract with multiple incentives formance should be weighted 507, schedule 107, and cost 407; i.e., the fee swing should be applicable as follows: Performance 507 ( ±$4.0m ) Cost 407 (±$3.2m ) Schedule 107 ( ±$0.8m ) ith these facts and assuming a simple linear relationship, the contractual arrange- ment can be portrayed graphically ( Figure 4). The Government must have carefully analyzed the matrix being established to ensure that the weights assigned do produce incentive com- Performance Incentive Provisions binations that promote the desired objectives. To structure the incentives properly so that they motivate the contractor towards the 2 desired goal, it is critical that the incentive values for each element (in this example, 507, s . 407, 107) be determined in light of and be o) consistent with actual technically feasible s “trade-off” possibilities among the several areas which are available to the contractor. The “trade-off” function represents the technical constraints on the contractor, i.e., the actual combinations of performance, cost, and sched-

ule which are possible or seem feasible to the « Schedule Incentive Provisions contractor. Without knowledge and considera- Z tion of the “trade-off function in structuring | j the incentives, there is no assurance that the g ^ desired Governmental objective ( increased z performance in this example) will be the same « objective selected by the contractor. To facili- INCENTIVE CONTRACTING 39

In terms of these variables, the contractor's the “trade-off” function, the optimum policy total fee may be calculated by the following for a contractor to follow can be determined formula: by mathematical programing techniques. The basic underlying assumption of incentive con- F t =.0SX — . 10666Y - .26666Z + 7° tracting is that the contractor will plan his provided that the values of X, Y. and Z are in activities so as to maximize his fee. That is, he the incentive region; i.e., X is between 0 and will select that combination of values of X, Y, 100 mph, Y is between $0 and $60 million, and and Z from all feasible combination alterna- Z is between 0 and 6 months. To illustrate, tives which would provide the maximum fee. note that if all incentive factors are on target For this illustration, the linear program solu- (X=50 mph, Y=$30m , Z= 3 months), the tion yields the following optimal solution: fee F t =$7m is as it should be. X = 50 mph To illustrate the impact of the actual Y=$0m “trade-off” function, it is first necessary to Z = 6 months determine what the “trade-off” possibilities Substituting these values in the formula: are. In an actual situation one possible method for making this determination is to request the F t — 08X - .10666Y — .26666Z + 7 contractor to provide forecast performance for F t =$9.4m several combinations of cost and schedule (with analysis by the Government as to rea- sonableness). For this illustration assume that the contractor submits the following: Figure 5. The “trade-off’’ function

Cost Schedule Performance $70m 21 950 24 975 27 1000 SIOOm 21 975 24 1000 27 1025 $130m 21 1000 24 1025 27 1050

Assuming that a linear combination of these parameters would be representative,-0 this three-way trade-off can be approximated in mathematical form as 1.2X — Y — 10Z < 0. See Figure 5. Given this mathematical representation of

O •The coefficient .08 represents the change in fee for each 6 mile per hour change in performance. (507, i 88 m _ nM

The coefficient .10666 represents the change in fee for each million-dollar change in cost. — O^' S8m = .10666 30 Tht* coefficient .26666 represents the change in fee for each month change in schedule. = .2ft666 The number 7 is a scale factor. 40 AIR UNIVERSITY REVIEW

Thus, the highest possible total fee the change to the fee (10%). The new target cost contractor could attain under this technical is now $110m and the new target fee is $8m “trade-off” condition would not be S15m as ( 7.3%). With regard to the incentive provi- assumed by the Government but only $9.4m . sions, the adjustments are commonly made in Further, and of utmost importance, is the fact either of two ways: that in this example the contractor did not a. Maintain the dollar relationships as ex- behave as desired or expected. Even though pressed in the original contract. Maximum performance was emphasized by structuring and minimum fees are adjusted upward by the the incentive fee 505? to performance, 40% to same dollar amount that was added to the cost, and 10% to schedule, the technical “trade- target fee ($1m ). Thus, the form of the share off" alternatives in combination with these in- line remains constant (shifted to the right centive fee percentages made it more advan- $10m and up SIm ), the fee swing is still tageous to the contractor to emphasize cost ± $8m , the range of incentive effectiveness is reduction. Note that performance does not still ±$20m , and the dollar amount of re- exceed target, whereas costs were underrun by ward/penalty for similar overruns and under- •$30m . runs from the new target cost is the same. It is significant that one of the alternatives Note this technique does not recognize any available to the contractor was X = 75, Y = 30, uncertainty associated with the change (ex- Z = 6, which appears to be more in keeping cept as negotiated in the new fee increment). with the 50%, 40%, 10% weighting factors, and (The original contract recognized ±$20m un- was not selected because the associated fee certainty, whereas the new arrangement pre- sumes 100 per cent confidence in the cost would be only $8.2m . In other words, when faced with the alternative of increasing per- associated with the change except as the formance by 25 mph or reducing cost by agreed-to weighting might affect them.) b. Maintain percentage relationships of the $30m , the contractor was motivated by the original contract and disregard the dollar re- fee structure to choose the latter. lationships. Using this technique the new tar- get cost and fee would be adjusted as in a, contract changes above, but the new maximum fee would be adjusted to $16.5m (15%) rather than to $16m Last, it is necessary to review another dif- and minimum fee to — $1.1m rather than to 0. ficult area of contracting even without incen- The contractor fee swing becomes -P$8.5m tive provisions; i.e., contract changes. Incen- and — $9.1m (rather than staying ± $ 8 m ). tive provisions compound the problem as the This illustrates that, by incorporating the impact must be determined not only on target change so as to maintain percentage relation- costs and profit (fee) but also on sharing ships, you increase the range of incentive ef- formula and ceiling price or maximum-mini- fectiveness which not only recognizes uncer- mum fee. Procurement regulations state, “It tainty associated with the change but also is essential that there be explicit agreement increases the uncertainty range of the old or between the Government and the Contractor unchanged portion of the contract. Unfortu- as to the effect on performance of contract nately, it accomplishes this in an arbitrary and changes.’ -1 Ideally, the negotiation of changes not necessarily rational manner. should not allow either party to reprice those While neither technique is completely elements of the work unchanged. Further, satisfactory, the constant dollar approach has when incentive provisions are involved, the some apparent advantage because it is easier changes must be handled in a manner which to administer and because it merely shifts causes the least possible disruption of the rather than changes the original share pattern. original incentive matrix. E.g., in the last illus- The constant percentage technique on the tration, assume that as the contractor pro- other hand changes the original incentive pat- gressed, it was necessary to negotiate a $10m tern for all incentive elements and the trade- change which resulted in a negotiated 4- $1m off ratios regardless of whether or not they INCENTIVE CONTRACTING 41 were actually affected by the change in tracts to do considerably more research re- work.22 garding what these writers have identified as The concept and objectives of incentive the technical “trade-off” function as well as contracting appear to have merit. It would how to administer contract changes effectively seem most profitable for users of such eon- and accurately. Air Force Institute of Technology

Notes 1. J. O. Wright. Chairman of the Board, Thompson Ramo 12. See Armed Services Procurement Regulation, Section Wooldridge. Inc., in a speech at a cost-reduction symposium III, paragraph 808.4. for the weighted ranges by individual sponsored jointly by the National Securities Industrial Associa- profit factor; e.g., direct material purchases have a range of tion and the Department of Defense in June 1961. 1 to 5% whereas engineering labor has a range of 9 to 15%. 2. Cost reimbursement contracts with percentage-of-cost Again, this effort may be too conservative to achieve the de- fees (CPPCi. first used during World War II, gave contractors sired result. an incentive to increase costs and hence fees. The cost-plus- 13. See DOD Directive 5126.38, 1 Aug 1963; DOD In- fixed-fee contract, successor to the now-prohibited CPPC con- struction 7700.4. 8 Aug 1963; and DOD “Guide to the Evalua- tract, does not provide incentive to the contractor to operate tion of the Performance of Major Procurement Contractors," more efficiently. 26 July 1963. 3. The War Department included a performance incentive 14. See Armed Services Procurement Regulation, Section feature in the 1907 Wright brothers’ contract. They were to be III, paragraph 405.4(b), dated 6 March 1964. rewarded by a 10 per cent bonus for each mile per hour the 15. Department of Defense, Assistant Secretary of Defense plane would perform in excess of 40 miles per hour. ( I&L), “Advanced Incentive Contract Workshop,” Part V, p. 49. 4. Arthur D. Little, Inc., as quoted by I. Kalikow, Federal 16. Reference the current DOD Directive 3200.9, 26 Feb Government accountant, in "Defense Incentive Contracting,” 1964, and Air Force emphasis on the definition phase (e.g., The Federal Accountant, March 1963, p. 68. AFR 375 series) for constructive efforts in this area. 5. Department of Defense, Assistant Secretary of Defense 17. Early delivery might well be desirable when the item (I&L), “Advanced Incentive Work Shop," 1963, Part I, p. 1. is independent of other systems or is on the critical path for 6. The Defense Department, according to a recent article development of an overall system. (Associated Press in Dayton Daily News, 31 December 1963), 18. Within context of overall objectives of the Government cited statistics showing that the trend of defense contract profit so that the incentive matrix developed causes the contractor to margins (on costs) is downward, in 1956 the average profit make decisions and perform in support of these objectives. being 6.3%, in 1962 3.1%. Reference subsequent discussions concerning this problem. 7. “Rickover Assails Undue Profits,” Dayton Daily Sews, 30 December 1963. 19. Performance could have more than one element set up S. See Armed Services Procurement Regulation, Section for incentive. Recognize, however, that the more elements in- III. paragraph lOlAai 1). p. 328; paragraph 405.4(a). p. 334 volved the more complex the matrix and contract administration. (and paragraph 40o.4(c), Limitations, for applicability of maxi- 20. The actual form of the “trade-off” function is most mum fee limitations to CPIF contracts); and paragraph 3, certainly nonlinear; however, over a limited range of values, a 405.5c( 2), p. 336, respectively. Regarding restrictions on fee, linear representation may be a reasonable approximation. For these can, in some instances, rule out any effective incentive further discussion, see Chapter 9 of The Acquisition Process by being incorporated, and thus, these writers believe, the Govern- Peck and Scherer. ment should seriously study the feasibility of relaxing such 21. See ASPR 3-407.2(a)(4). restrictions. 22. Headquarters USAF, in a letter, “DOD Incentive 9. Department of Defense, Assistant Secretary of Defense Contracting Guide, AFP 70-1-5,” dated 8 January 1956, states: (I&L\ "Advanced Incentive Work Shop,” 1963, Part I, p. 17. “. . . It should be recognized that either method can be used 10. Ibid., p. 16. and the determination as to which is preferred will depend on 11. For example, for one negotiation, an airframe manu- the specific procurement situation. Each has advantages and facturer submitted 45 pages on profits. disadvantages. . . .” ECONOMIC ASPECTS OF MILITARY ASSISTANCE

C aptain D ouglas N. J ones

H ILE IT IS generally not good the four basic components of the Foreign practice to begin an article on a Assistance Program. More obviously economic Wnegative note, two disclaimers are in aspect are the sister programs: Supporting in order: first, that in suggesting some of the Assistance, which consists of raw materials, positive economic aspects of military assist- commodities, machinery, and tools, plus finan- ance this paper should not be viewed as an cial assistance to support the defense produc- apology for military aid. To the economist the tion of our allies; Technical Assistance, which high "opportunity cost” of such assistance is consists primarily of technical cooperation well recognized and admitted to if the com- with developing countries; and Development parison is with the food, clothing, and shelter Loan Fund, designed to strengthen friendly demands of a country and its population. The foreign countries by encouraging the develop- point here is that it is equally erroneous to ment of their economies. As stated in the For- think of all military assistance as “down the eign Assistance Act of 1961 as amended, it is drain” in an economic sense. Second, even the purpose of this part (of the act) “...to within the current development of the recip- authorize measures in the common defense ient countries there is danger of attributing against internal and external aggression, in- too much to the military assistance program cluding the furnishing of military assistance where coincident progress is found. Again the upon request to friendly countries and inter- plea is that in its economic aspects the pro- national organizations.”1 This involves furnish- gram should be viewed as one significant form ing eligible military forces with the end items, of annual aid. Specifically stated, our question spare parts, and peculiar supporting, training, is, "How far does foreign economic aid under- and other equipment that will enable them to state total economic aid through the spillover resist external aggression and maintain in- economic effects of military aid?” ternal security. Consideration will be given to objectives The concept of a strength around the pe- and alternatives to military assistance, some riphery of the Communist world created and history and present status of the program, maintained by joining the capabilities of our- direction and magnitude, and issues and im- selves and our allies is central to the whole plications. security program. W e spend $51 billion annu- ally on our military establishment and have averaged about $1.4 billion on the Military Objectives and Alternatives Assistance Program ( ma p) during the past ten Military assistance is, of course, one of years. For this latter expenditure we gain men. ECONOMIC ASPECTS OF MILITARY ASSISTANCE 43

equipment, and bases from the recipient coun- (3) Military Assistance Advisory Groups try. These are estimated as 4 million men in ( maag ) are answerable to unified commands ground forces, 1900 combat naval vessels, and of the areas as the established channel to some S000 operational aircraft.- And since we dod; (4 ) military assistance plans are 5-year could not or would not likely supply all these time-phased and reviewed annually; and (5) ourselves and deploy them around the world, the Agency for International Development is the alternative to map would be some form of made responsible for coordinating map with the “Fortress America concept. both economic assistance and foreign policy. The questions of whether or not there has in fact been a shift from military to economic History and Present Status aid, the regional appropriateness of each, and the meaningfulness of the distinction are The current program can be traced back treated in the following section. to the transfer of 50 destroyers to Great Britain in 1940 and the Lend-Lease Act in 1941.3 Direction and Magnitude Large sums are not new to such programs. By the end of World War II the dollar value of Since World War II the United States lend-lease deliveries had exceeded $48 billion; has extended about $100 billion in foreign as- between 1947 and 1950 Congress had allotted sistance to over 100 countries and regional more than $645 million for Greek-Turkish aid; groups. Of this amount, the figures are over the Marshall Plan, which had significant mili- two to one on the side of economic aid. Even tary aid provisions, amounted to $12.5 billion in areas where military operations have from 1948 to 1951/ Added to this should be erupted dramatically to public notice and one expenditures for the period on various multi- might expect military expenditures to have lateral and bilateral arrangements, e.g., NATO predominated, this has many times not been and the Berlin Blockade. the case. And while it seems a good rule of Still at the heart of the present program thumb to say that our emphasis should shift are the three key provisions of the Mutual as the threat and our views of the threat shift, Defense Act of 1949—grant aid. direct trans- this is not to say that the two are mutually fers, and technical help. The Mutual Security exclusive. Economic aid “versus” military aid Act of 1951, understandably emphasizing the is largely a false issue—rather they generally military aid over the economic, consolidated support one another. First, even in cases where all the various aid ventures from the several economic development is judged to be a key Government agencies into a single Mutual objective, it is not always clear that a given Security Program and placed it under the D e- amount of economic aid is more conducive partment of State. Then in 1954 the act was to economic growth than an equivalent amended to divide the economic and military amount of military aid. This is especially so portions of foreign aid between the Depart- for some Latin-Ameriean countries, for ex- ments of State and Defense. Finally, the For- ample, where assistance for internal security eign Assistance Act of 1961, as amended in is predicated on the grounds that the military 1962 and 1963, replaced all others. Key provi- has an essential role as a stabilizing influence sions of these acts (together with an executive in these countries. Second, it can be argued order and administrative evolution) that make that in results—if not in labels—the two are al- for better control but more flexibility are (1) most interchangeable in that military aid will military assistance may be provided not only release resources and funds of the recipient by loan, grant, or sale but also by lease, ex- country for diversion into other uses, and vice change, or any other means; (2 ) the Joint versa. Finally, and our chief concern in this Chiefs of Staff are charged with continuous paper, direct spillover effects from military as- review in terms of global strategy- and security; sistance—like infrastructure expenditures and 44 AIR UNIVERSITY REVIEW civic-action programs, like road and harbor Military Assistance Program construction, communications and sanitation Comparison by Region facilities—clearly have significant economic as- pects.

The hypothesis that there is some kind of FY 1964* FY 1963 FY 1962 FY 1961 inverse relationship between military and eco- Europe 16% 15% 20% 28% nomic aid is dismissed by the Draper Report Africa 2 221 through its findings in 24 countries studied Near East and where the economic growth rate was less than South Asia 31 28 22 23 2 per cent.0 Ten of these countries showed a high rate of military investment, two a mod- Far East 44 44 46 41 erate rate, seven scarcely any military growth, Latin America 6 5 4 3 and five a zero or a negative military growth Nonregional 1 654 rate. Thus only 10 of the 24 countries lend any ’ (estimated) support to the argument at all. Further it must be remembered that economic growth itself Note: Percentages may not total 100 because of rounding. does not necessarily preclude Communist in- Source: Proposed Mutual Defense and Assistance Programs, cursions—witness the strength of the parties FY 1965, Summary Presentation to the Congress, Govern- in France, Italy, and Cuba. ment Printing Office, 1964, p. 206. At another point the Draper Report con- cludes that the matter of emphasis between economic and military aid should be dictated East, and Latin America increasingly heat up. by whether strategy appropriate to the coun- Military assistance reached a peak of $5.7 try involved is long-run or short-run in nature, billion at the time of the Korean War and the whether of the cold-war strategy or hot-war nato buildup. The request for $1.0 billion for variety. Thus if the threat is primarily exter- fy 1965 is close to the average appropriation nal, as in the case of countries on the Soviet that has prevailed since fy 1959, i.e., $1.4 periphery, military assistance would predomi- billion. The cost of military assistance thus nate, while economic aid may be more ap- represents approximately 2 per cent of the propriate to countries like the emerging Af- amount required annually to support the U.S. rican nations which are remote from imme- military establishment. diate military danger but prone to internal Finally it should be pointed out that for threat.7 many countries their defense effort is con- However this may be, it is clear that the sequential relative to their gross national prod- Military Assistance Program itself has been uct ( gnp) and large compared to other com- cpiite responsive in magnitude and direction mon measures of economic magnitudes. A 1961 to the thermometer of international tensions. rand Corporation study of seven countries Not surprisingly, the bulk of U.S. military aid over the 1958-59 period showed the median has gone to the world’s trouble spots of the proportion between defense and gnp to be 5.2 postwar era. The most casual newspaper read- per cent; between defense and gross invest- er would probably arrange the order about ment, 34.3 per cent; between defense and as the evidence indicates in the accompanying Government expenditure, 53.3 per cent; and table. Note that over the period about one half between defense and Government investment our aid has gone to Europe, one third to the 81.7 per cent.8 In economic terms the question Far East, one sixth to the Near East and South is posed, How can military assistance, and the Asia, and tiny amounts to Africa and Latin structure of defense forces and budgets in America. Note too that even in the short underdeveloped countries, be modified so as period presented (1961-64) the trends are to yield approximately equivalent military ef- suggested as Europe stabilizes and picks up fectiveness and yet generate substantially im- more of its own tab while South Asia, the Far proved economic and political effects? ECONOMIC ASPECTS OF MILITARY ASSISTANCE 45

The study sets out the principle that If it is felt that military assistance is not where the primary criterion—in this case, mili- particularly well suited to meeting erupting tary effectiveness—does not indicate obvious immediate crises, it is all the more clear that preference for a particular alternative pro- any program must be carefully linked with the gram, the choice should be made on secondary overall pattern of U.S. strategic policy for that criteria. This might be a consideration of eco- country. nomic performance through a comparison of Third, while there is general agreement the effects of alternative military programs on that economic assistance must accompany and the operating costs of the defense establish- be integrated with military aid (especially in ment, on public capital formation, and on the the case of underdeveloped countries), the generation of skills in the recipient country. tests for military participation in economic de- Here one would choose the program with the velopment are not so clear. Three suggested best ancillary economic developmental effects guidelines might be where participation (1) on the basis of the differential opportunities. does not unduly detract from the military task, In the concluding section I shall take up (2) is not contrary to the long-run develop- some of the issues that are still loose in the ment of institutions of popular preference, and Military Assistance Program and cite some of (3) is clearly in the public’s interest, as op- their implications. posed to that of any special-interest groups. Examples of productive participation of local military people in civic-action programs are Issu es an d Im p lica tio n s myriad. These have typically come in the areas of agriculture, industry and mining, Finding the appropriate focus of attention transportation and communications, sanitation along the spectrum of resistance to external and housing, labor skills, and public adminis- and internal threats presents the policy- tration. They include crop dusting by the Air maker with a serious dilemma. If too many Force in El Salvador and veterinary service resources are devoted to countering the by the Army in Argentina. external threat, the recipient country be- The technical training of foreign military comes vulnerable to internal instability, e.g., personnel is frequently cited as a major aspect Korea. If too much is devoted to the internal of the Military Assistance Program. Approxi- threat at the neglect of the external, adven- mately 175,000 foreign nationals have been turous grabs may be invited, e.g., India. And trained in the U.S. under the program, and if the resources of resistance are spread too about 60,000 more have been trained overseas thin in combination, exploitation of the vulner- at LhS. installations.9 Still others have been abilities can be expected, e.g., Laos. Further- trained in their own countries by U.S. teams. more, internal stability and order may come Besides their hoped-for political role in stabiliz- at considerable cost, as is sometimes alleged ing internal upheavals, their spillover skills in the case of Latin America. and educational and technical competence can Establishing helpful standards of judg- be viewed as a substantial input to the human ment for deciding what countries should re- capital of their countries. ceive military aid ( and how much) is a second Finally, passing mention should be made difficult task. Some argue for an annual, ad of the relation of map to the balance-of-pay- hoc, case-by-case analysis of the distribution of ments problems currently experienced. Testi- the cost burden, while others contend that the fying on the Foreign Assistance Act of 1962, uncertainty of assistance inherent in such an General Lyman Lemnitzer, Chairman of the approach is destructive to the posture of the Joints Chiefs of Staff, said, “There is a collat- recipient country. Presently annual reviews of eral benefit of our Military Assistance Pro- longer-range aid plans, with significant down- gram which is not generally appreciated. Our ward modifications largely limited to the end Military Assistance Program does not result years, seem to satisfy the latter objection. in a flow of gold from the U.S. The contrary 46 AIR UNIVERSITY REVIEW

is true.’ 10 The $335 million of map funds spent It is likely that the above issues, and outside the U.S. in 1962 was offset by $800 others, will continue to be debated as with all million received that same year from the sale Governmental programs where the costs are of military equipment under the program. The easy to determine and the returns are not. It receipts for 1963 were reported as $600 mil- is equally likely that the Military Assistance lion. Of course the shifts from foreign buying Program will continue at about its present size to U.S. buying and from grants to sales ( cred- (not to say direction), barring either a dra- its) were the key changes that reduced the matic change in the long-term international domestic trade implications of the Military climate or a profound shift in U.S. defense Assistance Program. strategy. United States Air Force Academy

Notes (6th ed.; Washington, D.C.: Headquarters USAF, 1962), pp. 8-16. 1. Foreign Assistance Act of 1961, S. 1983, Public Law 6. Supplement to the Composite Report of the President’s 87-195, 87th Congress, 4 September 1961, Part 11, Chapter 1, Committee to Study the United States Military Assistance Pro- p. 10. gram ("Draper Report), II (Washington, D.C.: Government 2. Foreign Assistance Act of 1963, S. 1276, Hearings be- Printing Office, 1959), Annex C, p. 62. fore the Committee on Foreign Relations, U.S. Senate, 88th, 1st, 7. Ibid., pp. 36-3 7 . It is recognized of course, as men- June-July 1963, p. 181. This does not of course imply any direct tioned, that “stability aid” may be required in the form of involvement of these forces with the U.S., except to the degree military assistance even in nonperipheral countries as an element that they are firmly committed with us in bilateral or multilateral of order and authority. Also the whole idea of counterinsurgency defense agreements. and its evolving requirements may severely upset this neat dis- 3. The historian would likely suggest earlier spillovers of tinction. military/economic development in the U.S., e.g., West Point and 8. C. Wolf, Defense and Development in Less Developed the Corps of Engineers, exploration of the West, Navy geodetic Countries (California: RAND P 2291-1. 1961), p. 1. The coun- surveys, air charting and graphic services, flood control, canals, tries studied were Burma, India, Indonesia, Korea, Pakistan, and railroad activities. Philippines, and Viet Nam. 4. Military Assistance Program: Information and Guidance 9. Foreign Assistance Act of 1963, p. 12. (5th ed.; Washington, D .C.: Headquarters USAF, January 10. Foreign Assistance Act of 1.962, pp. 66-67. See also 1960), pp. 1-4. Secretary McNamara’s testimony, Foreign Assistance Act of 5. Military Assistance Program: Information and Guidance 1963, pp. 181-82. V 42

The Science Frontier

RESEARCH IN COMPUTER SCIENCES

D r . H ans H. Zsc h ir n t

VIDENCE of the explosive expansion of lance and control of traffic in the air and on the E data-processing techniques is all about us. ground. The result is an almost endless list of Our pay checks and mortgage statements, our driv- applications, actual and potential, of data process- er’s licenses and magazine subscription forms are ing, not to mention the vast amounts of technical growing more and more alike in appearance—rec- and scientific computation that require computers tangular cards with one corner missing and gar- of ever increasing capabilities. nished with mysterious, likewise rectangular, holes. The armed forces, and particularly the Air These cards are destined to tell some machine Force, are making extensive use of military varia- something about us that we are not always sure tions of many of the applications of automated we want the machine (or anybody, for that matter) data processing mentioned so far. But they add to know. That these holes have a particular ten- to them their own specifically military require- dency to cluster in that part of the card where it ments. It would lead too far to examine these says, “Sign here,” may account in part for some special requirements in detail, but some general popular resentment against automatic data proc- properties of the military environment are evident. essing. For example, there is the sheer size of the opera- All operations that require large-scale record tion to be controlled (as in global logistics opera- keeping and manipulation, such as banking, cen- tions) and the need for processing in real time, sus, tax, libraries, transport scheduling and reser- i.e., the need for the data-processing operation to vations, and other logistics operations, are rapidly keep pace with the unfolding processes that it on their way to becoming fully automated. Added evaluates or controls. to this are control functions such as space opera- Furthermore, when an information-processing tions, process control in manufacturing, surveil- system is conceived and designed, it is often not 48 AIR UNIVERSITY REVIEW possible to predict completely the data environ- chines envisioned by Turing in the 1930’s and von ment in which it will have to operate. That may Neumann in the 1940’s. They are sequential in greatly depend on some future strategic situation. operation, have a stored program, and possess We may also not be able to describe completely memory. the specific goals of the system in actual use. This The information on which the computer op- calls for a high degree of operational flexibility. In erates is encoded in binary digital form as addition, many military systems operate in the sequences of zeros and ones, most frequently sig- presence of inevitable uncertainty. We hope they nified by the presence or absence of an electric will be capable of helping us make the best of voltage. This coded information is usually ar- that situation. Further, many military systems are ranged in groups of a fixed number of digits, often intimate composites of decision-making and oper- called words, and sometimes in subgroups corre- ating people and evaluating, computing, digesting, sponding to characters. Let us remember that four displaying, and decision-aiding machinery. The binary digits (bits) are necessary to encode a presence of the man/machine interface—as well as decimal digit and five to seven, depending on the the vital importance of its smooth functioning—is size of the alphabet, to encode alphanumeric char- one of the most crucial aspects of information- acters. processing systems in the military domain. Other binary sequences, following the same There are also special physical requirements word organization, are characterized as “instruc- on the equipment, such as transportability, need tions,” i.e.. they are interpreted by the machine as for extreme miniaturization or radiation- or heat- steering signals which cause certain discrete resistance. operations to be performed on the data whose Research in computer sciences, or, better, location in memory is indicated in the instruction information sciences, has thus become a vital part word. Thus the instruction code is the internal of our scientific picture. It is also evident that the “language” of the machine. The sequence of in- armed forces have a considerable stake of their structions, describing the complete computing own in this research. process desired, is called the program. Both In this article we shall examine, without any program and data, externally furnished or in- claim to completeness, some of the more pressing ternally generated and needed for future refer- problems that presently exist in information sci- ence, are stored in the memory in word locations ences. In doing so, we shall mostly use a rather bearing numbers (addresses) and can be re- utilitarian approach, such as asking, “What do we trieved, changed, transferred, or discarded when have to do about information-processing systems called for in an instruction. Finally, there is input so that the computers will do the most for us?” gear for the introduction of data, programs, and ad This attitude will not save us from some thought hoc operating instructions, and there is output transcending the confines of computer technology. gear for the communication of results to the user. To give a sensible answer, we have to examine the The speed of internal processing is one of the entire complex of computers, people, the environ- striking features of digital machines. Two mil- ment they live in, and the nature and mechanisms lionths of a second per instruction is rather com- of their interactions—in other words, a goodly por- mon in modern machines, and further increases in tion of the realm of cybernetics. Nevertheless we speed by orders of magnitude are just over the shall take a somewhat restricted view based on horizon. Desirable as speed may be—and there are some applicational viewpoint. But in the back- important needs for further speed increases—it ground lurks uncertainty about the fundamentals. appears, however, that speed is not our biggest problem at the moment and certainly not for the more distant future. some properties of digital computers This broad-brush sketch of a typical computer Common to all automatic information-process- suggests a number of properties which have given ing systems is the fact that they are centered about rise to important problems for research in informa- a digital computer. All digital computers in opera- tion sciences, some of recent emergence and some tion today are direct-line descendants of the ma- of long standing. THE SCIENCE FRONTIER 49

Because a stored program can be altered at contents of its memories plus complete information will or replaced by a completely different sequence as to which of its internal switching devices that of instructions, the machine can perform an in- are the carriers of the logic manipulations are “on” finite variety of tasks without change of its physical or “off” at the moment.) In a machine that is ex- configuration. This lends to digital computers a pected to perform a useful and describable task, universality unparalleled in other physical tools. this sequence of states corresponds to the decom- The internal communications of the machine occur position of the task into machine actions or, in other in the abstracted form of binary symbols, regard- words, constitutes the description of the task in less of their nature as data, instructions, addresses, terms of machine language. Nothing in this se- or internal housekeeping signals. The latter three quence can be left to chance. In particular, there can therefore be operated on, if desired, in the must not be any ambiguity at any given moment as same manner as the data. Thus, the machine can to which way in the universe of states the machine modify its own program and automatically rear- will travel next. This is true whether the problem range its operational structure; it is, in short, capa- is explicitly programed in the machine or whether ble of manipulating abstract symbols in any way the machine is caused to generate its own proce- that can be described in terms of mathematical dures by being programed in terms of “procedures logic. From this, the term “logic machine” derives to produce procedures,” etc. its justification; a whole branch of mathematics, In the ultimate sense, the dictum that “you automata theory, concerns itself with the proper- only get out what you put in” holds true. Even if ties and behavior of generalized computers and the behavior of a machine* as observed at its out- uses for its mathematical tools concepts and meth- put seems to be the result of considerable mental ods touching closely, on occasion, upon the very activity of its own, it is always retraceable to the fundamentals of mathematics. preconceived layout of the machine and can thus Thus, our machine is capable of much more be explained, in principle at least, in terms of than just rapidly and accurately executing numeri- straightforward switching theory. There are, con- cal calculations or keeping books on all sorts of trary to popular belief (and, I am afraid, to the events involving large amounts of data. Its symbol- subconscious belief of some of the cognoscenti), no manipulating capability enables us to write pro- miracles in this business. grams that generate other programs (a capability It is evident, however, that the correspond- used in compilers, which we shall discuss later) ence between a problem as we see it in terms of and to modify programs or parameters conditional our human frame of reference and the way it on past results. If such modifications lead to an presents itself in machine terms is extremely improvement in performance, this might be con- tenuous. The user needs an interpreter to converse strued as a certain rudimentary ability of the ma- with this type of machine. Thus has arisen the chine to “learn. Since logic operations in the ma- priesthood of programmers and analysts and, for chine also include operations of comparison and their benefit, the proliferation of intermediate lan- merger of similar symbols according to certain guages to serve as buffers between human means criteria of similarity, one might infer a basic capa- of communication and the idiom of the machines. bility of forming concepts, “recognizing,” and If machines are really intended to be extensions of making decisions—all aspects of behavior akin to our mental capabilities, the narrowing—or better, human intelligence and implying the great poten- bridging—of this gap is a necessity that has become tial usefulness of machines as extensions, “ampli- universally recognized. fiers,” and helpmates of our own mental processes. A second point is that, for all practical pur- On the negative side of the balance, there is poses, we have high expectations of the infinite the fact that as logic mechanisms, machines operate number of tasks that machines can do equally well in a rigidly deterministic manner in the sense that or better than humans. Many current uses of com- each state of the machine is completely determined puters give a hint as to what is possible. On the by the sequence of previous states and the latest 0 By "machine” we mean either the machine as a physical input. (The “state” of a machine is the ensemble structure or a machine plus its program. One machine executing two different programs may thus be considered, for the purpose of the values of all its internal variables, i.e., the of this article, as two machines. 50 AIR UNIVERSITY REVIEW

other hand, disagreement is widespread on what badly enough, it can be realized, although at a the limitations of machine performance are. Opin- price. Thus we are turning more and more to an ions, often supported by vigorous argument, range examination of new concepts. from “Just another appliance” to “Fifty years from To a certain degree, our present problems now computers will have a symposium on ‘Can peo- with computers and their use in a man/computer ple think?’ ” There is ample reason to assume that environment are a direct and inevitable conse- the truth is intermediate between the two extremes. quence of the evolutionary history of computers. But the border between the proper and effective As we remarked before, the early creators of logic- use of computers and the efficient and human use machines were perfectly conscious that their basic of humans is disputable. Not only are we uncertain potentialities reach far beyond the immediate pur- about the limits of capabilities of future computers pose of high-speed programed numerical calcula- but we are also not too sure about the basic tion. At the time, however, the shortcomings of the strengths and weaknesses of humans regarding contemporary electronic technique put the burden data-proeessing tasks in the broadest sense. of designing viable machinery on the electronic en- No matter what the division of assignments gineers. Many computer designs in present use are between man and machine may be, there is in- the result of electronic engineers’ designing ma- evitably an interface where easy and efficient com- chines to the best of their not-inconsiderable knowl- munication between the two has to occur in order edge and understanding. The users, mathematicians to make Man-Computer Symbiosis (Licklider) and programmers, were not always very helpful. happy and productive. A major portion of present The electronic engineers handed the finished prod- computer research is therefore concerned with uct to their customers and said: “Here is the best man/machine communications. Considering the we can do; program it, and good luck.” As a con- present state of computer technology and use and sequence of this evolution, the overwhelming how it came into being, one might say that much majority of computers in current use are, basically, of the effort in the last decade was, first of all, to tailored to be highly efficient sequential numerical make computers work, and work well and reliably. calculators. Only recently more attention is being The operating speeds and the reliability of com- given to the increasing use in nonnumerieal tasks puters have increased dramatically, while at the by provision of a larger repertoire of logical in- same time problems of maintenance and space and structions and a capability of parallel operations power requirements have been substantially re- and features to foster a rapprochement between duced. Solutions—at least preliminary ones—to the user and machine. problem of man/machine communications in the Current and future research in the information form of elaborate programing systems and occa- sciences is therefore likely to be much less domi- sional techniques of direct man/machine interac- nated by devices and electronic techniques than it tions have been devised and put into use. The was in the past. It is much more vitally concerned existing amount of background research on prob- with the interrelation of man and machine. Ques- lems of artificial intelligence is staggering, if meas- tions that should be in the foreground of our re- ured in terms of printed pages published. Nu- search go like this: merous works have been produced on the problems (1) In a man/computer system, how should of new and different machine organizations. New the work be divided between humans and ma- electronic techniques have been evolved, particu- chines? What can machines do better, and what larly in the field of microelectronics and thin-film can humans do better? integrated circuitry, which will make possible ma- (2) How can the communication gap between chine configurations that were fantastic only five man and machine be bridged so as to minimize the years ago. The emergence of these new techniques existing difficulties at the man/machine interface? contributes a problem—namely, how to use them (3) Is it necessary to make radical changes in to the best advantage—and a relief at the same the concept and organization of our computers in time. Contrary to the past decade and a half, we order to solve the first two problems, and how can can be reasonably sure that, if we need and desire newly emerging electronic techniques be used to a certain operational characteristic in a machine help realize those new machines? THE SCIENCE FRONTIER 51

I shall elaborate somewhat on these questions. near future. Promising signs have been seen lately; the field appears to become less clamorous, less glamorous, and more objectively matter-of-fact. Division of Assignm ents Between M an and. Machine problem solving and games Humans cannot match certain computer capa- bilities. Any repetitive task of numerical computa- Two aspects of machine-implemented “intelli- tion, bookkeeping, or data sorting that can be gence” (here taken only in the pragmatic sense of defined in terms of deterministic instruction se- a “cooperative attitude on the part of the ma- quences is definitely within the realm of com- chine”) are of particular interest, namely, the use petence of computers. Included are functions that of machines for problem solving and the machine are normally considered as requiring considerable implementation of cognitive processes. Typical for intelligence, if performed by humans. Bookkeepers the problem-solving aspect and of great continuing and newspaper layout and typesetting people are interest in research over a number of years are such certainly qualified as highly skilled workers, but tasks as theorem proving and the playing of games. computers do just as well, if not better. On the The purpose here is not so much to create a dis- other hand, computers are often inferior even to cipline of “mechanized mathematics” (Wang) or small children in performing tasks requiring recog- just to prove that machines can be as good chess nition or concept formation, although there are or checkers players as people, but rather that a strong indications of a considerable potential for generalized solution of such problems would be machines in that regard. For this reason a con- of staggering significance for problem-solving pro- tinuous appraisal of the relative capabilities of men cedures in strategy, economics, and social sciences and machines is necessary to make the best current that presently overwhelm us humans by their com- use of man machine systems. plexity'. The problems mentioned have in common some aspects of artificial intelligence that, in principle, the)’ are amenable to an algorith- mic0 solution. A barrier that machines run up It is not surprising that the field of artificial against with algorithmic processes when perform- intelligence, concerned with automata displaying ing any but the most trivial tasks is that an exhaus- behavior that in human terms could be described tive search ol all possible proofs or all possible plays as requiring a certain amount of intelligent think- is impossible even for the fastest and largest ma- ing, has been very much in the foreground of com- chine, not to mention the utter lack of elegance of puter reseach and will certainly continue to be of such a brute-force procedure. Instead, we are look- foremost importance. In fact, recent bibliographies ing for approaches to successful solutions of prob- on artificial intelligence—e.g., the excellent an- lems without exhaustive search of the entire solu- notated bibliography by Marvin Minsky—contain tion space. These approaches, justified mainly by the about 500 items written between 1959 and 1961. fact that in similar situations they were successful, Another bibliography of somewhat broader scope, are called heuristics. They are extensively used in covering the years from 1946 to 1959, contains problem-solving cases using heuristic programing. about 4000 items. In part, this prodigious prolifera- In such cases, certain heuristics, for their experience- tion is misleading as to its real content. The diffi- proved usefulness, are built into the program and culty of defining “intelligence” in a commonly ac- are selected if the type of problem at hand seems ceptable way is great, and we will not attempt to to warrant their use. To determine what the type go into any depth in that respect. Further, since— of the problem is, the machine has to perform an luckily—authors with varying backgrounds have act of classification, to recognize whether the pat- been fascinated by the subject, it turns out that tern of the problem calls for this or that type of there are many different ways to say the same thing. procedure. Consolidation and unification in this respect, to- At present, by and large, the intelligent selec- gether with some weeding out of the obvious or •Algorithm. Some special process of solving a certain type redundant, will certainly take place in the very of problem. FIRST DEMONSTRATION 01 tion of avenues of approach and the machine-aided TO PROVE. )P)Q@)RP)RQ playing of hunches” seem mainly a human prerog-

COMMANDS. I), I) ative, although increasing mechanization of even this function is definitely in the cards. Presently, at I) I) least the design and programing of the heuristic procedures to be used are tasks performed by hu- PROOF TO DATE mans, but the selection of the procedure called for *P by the problem at hand may be mechanized. The *Q Mr* next step, of course, is to cause the machine itself (3)RP.)RQ )Q. .(5>)RP. )RQ to “invent” useful procedures or to improve existing )P*)Q..«3)RP.)RQ ones according to their past success or failure. In the area of problem solving, as in other in- stances of simulation of intelligent behavior by COMMANDS. 0A8 05 machines, the enthusiasm of the research worker THE CURRENT COMMAND IS NOT WELL-FORMED. has sometimes tended to get in the way of rapid COMMANDS. 06, 05 practical progress. What I am referring to is the ) RP )RQ attempt to mechanize an entire reasoning process and the complete refusal to accept the available PROOF TO DATE aid of the human operator. I hasten to declare that *P for research purposes, where the objective is to find *Q*** out how completely a mental process is replicated )RQ by a computer, this procedure is, of course, the ) RP only possible one. In cases of practical interest, «a)RP. )RQ )Q..?)RP.)RQ however, the machine alone tends to fail even at )F*)Q. .(5>)RP.)RQ moderately complicated tasks. It is here that the collaboration of a man/machine team has the great COMMANDS. I), I) promise for the near future. Particularly in complex I) situations—and decision-making systems of military I) interest are inevitably very complex—this approach PROOF COMPLETE ( INTUITI ON ISTIC) has the greatest promise and stands in the fore- 01 *P ground of interest. 02 *Q 03 *R As a simple example of man/computer cooper- ation, Figure 1 shows a computer print-out of a 06 theorem-proving task in propositional calculus. 05 )RQ 06 *R Without explaining the problem or going into the 07 P 08 ) RP details of the notation, we shall only note that a sort 09 )RP. )RQ place. After accepting and acknowledging the statement of the problem, the machine “asks for COMMAND HISTORY. suggestions on how to proceed (“commands ). I),l). )RP,)RQ. Having received those, it goes through a sequence I),l). ot proof algorithms, then finds that a proof has not been achieved, and requests further suggestions. At this point we see that commands are not ac- cepted unquestioningly. The line the current Figure 1. Man/machine co- c o mman d IS N O T well- fo r med” is the machine’s operation in theorem-proving polite way of saying that the suggestion made was illegal. After correction of that error, the proof then proceeds to a successful finish. Many successful examples of man/computer teamwork are presently under development or in early operation. The ap- THE SCIENCE FRONTIER 53

plications covered include such diverse subjects as The general problem of pattern recognition or the machine production of technical drawings, the automatic classification can, then, be formulated as efficient layout of hospitals, and the flow-charting follows: The events or objects to be classified are of new computer designs. described by sets of measurements of distinctive features, frequently called descriptors, of the event. By suitable evaluation of these sets of meas- pattern recognition urements, it is desired to assign the unknown event In the center of discussions of artificial intelli- to one of a set of categories or classes. To do this, it gence are problems of recognition and classifica- is necessary to emphasize those features which tion, often referred to as “pattern recognition.” This tend to enhance the dissimilarities between mem- is rightfully so. The capability to deduce from a bers of different classes and the features most com- limited number of examples of a class of objects mon for members of the same class. At the same their common characteristics, to form a generalized time we will have to suppress features which are “concept” of this class, and, on the grounds of these more or less present to the same degree in all the concepts, to assign new unknown objects to their classes concerned and which therefore do not con- proper classes is certainly one of the striking tribute to the recognition process. achievements innate to human (and to a certain Basically, the development of an effective extent animal) mental activity. Formation of con- recognition process is critically dependent on the cepts, except where they are built in a priori (in procedure of assigning weights to measured fea- which case our so-called cognitive process would tures. These weights are usually, explicitly or im- be reduced to a simple deterministic sorting proce- plicitly, derived by examining a statistically signifi- dure), always implies a certain process of adapta- cant number of samples of objects belonging to tion or self-organization in the machine in question each class and assigning weights to the measured —hence the almost synonymous designation of sell- features according to their contribution to efficient organizing systems, or adaptive machines, or the classification. This process can be and has been more anthropomorphic attribution of a capability automated. The establishment of sets of weights to learn. constitutes a sort of “learning” or “training” phase If in this discussion we concentrate on pattern- for the machine not unlike the processes involved recognition procedures as prototype of adaptive/ in the training of a child or an animal. Let us take cognitive systems, we derive justification from a machine intended to read printed, upper-case the fact that command control or decision systems roman letters. Let us say our measurements to or almost any artificially intelligent system of rea- yield distinctive features measure such things as sonable complexity always has to ask as its primary area covered by black ink, length and location of and crucial question: What does all this mass of line elements forming the letters, curvature, loca- input data mean in terms of selection of appropriate tion and angle of intersection, and the like. By action from a limited set of possible courses? Recog- examining a number of samples of letters in dif- nizing the “pattern” underlying the input data of ferent type fonts, we (or the “learning” machine) the moment is the most important step to a solution. find that in an upper-case “A” the fact that it is Since the machine procedure operates on input composed of three straight-line elements two of data encoded in machine language and therefore which intersect near the top at an acute angle oc- divested of their physical implications, pattern- curs in most “As” and should be weighted heavily. recognition procedures using any given principle But some “A s are flat at the top. The feature that proceed regardless of the data source, be it the the “A” is composed of straight lines deserves not automatic reading of print, the interpretation of too much weight because too many other letters spoken words or of photographs, or possibly situa- have the same feature. In the “B” the presence of tion assessments pertaining to the field of eco- two closed loops, one on top of the other, is of nomics or military strategy. There are, of course, heavy weight. In all cases, the feature that the differences in complexity depending on the size character is printed black on white deserves zero of the data base, the number of alternatives of out- weight because it is present in all characters. This put choice, etc., but no distinctions in principle process of generating efficient sets of weights is that depend necessarily on the data source. quite central to recognition and decision processes. 54 AIR UNIVERSITY REVIEW

It can also be described in a number of different of mathematics does not restrict us to any specific ways, which fact accounts in large part for the number of coordinates, we thus plot the location bulk, diversity, and confusion of the literature on of an input event as a point in a multidimensional this subject. hyperspace,’ meaning a space having more di- Although it may be a little rough going, I will mensions than the conventional three that are open make an attempt to explain how this comes about. to our sensual perception. In the preceding example we have described the Weighting and summing procedures, as in the assignment of weights to descriptors as a rather case of matched filters, result, in this concept, in straightforward process of evaluating the statistics moving the points in hyperspace around sb as to of known samples in terms of their distinctive fea- encourage their clustering in a desired u'av or so tures. Note that we were careful not to state in as to produce a number representing a “degree of detail how this statistical evaluation is done; there fit" that may tell us to which class to assign the are numerous ways to skin that cat, and different event in question. The matched-filter procedure schemes use different ways. described earlier is actually, mathematically speak- So far we have talked about features in easily ing, nothing but computing the scalar product of comprehensible ways—in terms of line elements the “event vector” and the “weight vector.” In per- and intersections in the case of printed characters. forming this abstraction, we have gained much For those familiar with radar signal detection, let more than a certain convenience of mathematical us consider that problem as one in pattern recogni- manipulation. We have divested our inputs of tion. For a radar receiver, there exist only two physical significance without taking away any of classes of events: signals emanating from its own their information contents. We have, further, re- system, and meaningless interference. It has been duced the “filtering” or “weighting” concept from found that by analyzing the frequency spectrum a process of selective attenuation or emphasis to of the incoming signals and admitting those fre- a mathematical operation of transformation in quencies that are present in the original trans- hvperspace. mitted signal to a degree proportional to the degree While the physical constraints of the earlier they were present when transmitted, we achieve operations tend to confine our thinking to some- an optimal discrimination between signal and in- thing that is. somehow, physically feasible, the terference. This technique is known as “matched transformation concept leaves us perfectly free to filters." Look at it from the pattern-recognition do anything that is allowable under the rules of point of view: frequency components = distinctive our chosen mathematics. We can pile transforma- features; degree of admission of those frequencies tions upon transformations, make them linear or = weight. It seems that we can also describe our nonlinear to any degree—ours is the choice and the pattern-recognition scheme as a set of matched responsibility. With this generalization, finding a feature filters! decision procedure becomes a formal process de- Another view of the problem can be had by void of physical meaning. In the long run, it can introducing the rather abstract idea of patterns be assumed that such general procedures at the as vectors in multidimensional space. Let us ex- decision end of the game will become of high sig- amine this idea briefly as an example of consider- nificance, notwithstanding the temptation to make able mathematical abstraction and of the advan- the study of those things a vehicle of theorem- tages that an abstracted view can have. Assuming generating mental exercises without great rele- that we operate with only two descriptors, we can vance to the central problem. plot their magnitude along the x and y axes, re- spectively, of a graph. The point determined by the question of significant features those two coordinates would be a representation of one input object. If our descriptors are chosen So far we have assumed that the descriptors appropriately, we would hope that those objects derived from measurements on an event or an belonging to the same class would be close together object were sufficient and reasonably suitable for on our graph and those belonging to another class the classification process in question. In reality— eonvenientlv clustered elsewhere. Since the world and this is one of our major problems in producing THE SCIENCE FRONTIER 55

somewhat sophisticated pattern-recognition or de- as words No. 1 and 2, and so forth, thus illustrat- cision systems—this is pitifully not so. The reason ing the total independence of the decision process is not unawareness of the problem but lack of clues on the data source. as to what to do. To discover these clues is one In current successful procedures of recognition, of our major objects in information-processing re- one striking dissimilarity between existing machine search. In almost all the automatic classification procedures and human recognition seems to be systems that are operating successfully, the descrip- apparent. This is the much higher degree of ab- tors and the measurements leading to them are straction present in the human process. As an the results of human experience and some guess- example, let us look at an afcbl program con- work. or the plain outgrowth of the limit of what cerned with the solution, by machine, of aptitude the current techniques in electronic measurement tests based on geometric analogy. Such tests look can deliver. Fortunately, this is good enough, or like Figure 2. The question is: “Configuration a nearly so, in many instances. Printed type can be is to b as c is to which of the other configurations?” read and square pegs can be distinguished from The answer, in the example, is configuration 4. The round holes. In more complicated situations, the machine has consistently performed in these tests discovery of proper descriptors is, maybe, more somewhat better than the average high-school crucial than the decision procedure used. Our goal, graduate—so maybe this thing does not really test then, must be to discover descriptors that ade- intelligence. Anyway, the machine goes about the quately—and, hopefully, optimally—describe the job as follows: It lists in its memory, for each con- kind of events we are interested in for the purpose figuration, the coordinates of all points covered by of the specific classification we want to perform. the configuration; combines them into line ele- In order to develop the descriptors, we have to ments and lists those, including their end points consider not only the kind of events we expect to and curvature; determines, locates, and lists inter- deal with but also the goal of our classification sections and angles and which part of the figure is process. For example, if we are dealing with solid inside of which other part; and so forth—a tre- objects and want to categorize them as '“solid” or mendous number of minute facts. Then it deter- “hollow,” a system of feeler gauges might be a mines the transformation that has to be performed proper measuring device; if. however, our intended to convert a into b, and c into each of the other classification is by color, the color-blind gauges figures, and which of these latter transformations don’t do us any good. Unfortunately, in many of is most similar to the a-to-b one. It works, but it is the really interesting problems, present methods certainly not the way humans go about this task. of measurement and description may indeed be Figure 2 was shown to a number of people in our “color-blind” in some sense and account for fail- laboratory, and 80 per cent of them agreed with ure of the system. The attempt to mechanize a the intended solution. They were asked to describe system that observes its own performance and the reason for their decision. A typical answer was: changes its choice of descriptors, and the weight “In configuration a, a ‘clean, solid’ outline is inside assigned to them accordingly, is an important of something not quite so clean and solid; the branch of the research in artificial intelligence. same can be said of c and 4, respectively.” It Let me mention in passing that at the Air appears that, in solving the problem, humans tend Force Cambridge Research Laboratories ( afcrl) to use abstract and, at the same time, often quite an adaptive recognition system is under investiga- hazy descriptions. This method enables them to tion that generates property filters solely on the deal with situations not previously encountered grounds of the statistics of events unfolding in more efficiently than the machine and, it seems, time—somehow on the premise “I don’t know ex- more economically (note the striking difference in actly what I’m looking for but something must numbers of descriptors used). be there." Indeed it has had encouraging success An interesting approach to the development in discriminating heartbeats of different people and of “learning” machines is to conceive the learning different spoken words fed to it simultaneously and process as a genetic process. This works roughly intermixed. It has achieved separation of heart- as follows. Assume, to begin with, that one designs beats of person No. 1 and person No. 2, as well a machine capable of performing a certain task, a b c

1 2

Figure 2. A geometric analogy problem although quite inadequately. Now, we produce a sciences in general has indeed accumulated a number of replicas of this machine, but each rep- wealth of material about the strengths and weak- lica has small, random variations from the original. nesses of human performance. Added to this are Thus we have obtained a new “generation” of the results of practical experience of human per- machines with small “mutations.” From this gener- formance under stress—of particular importance in ation we pick those that perform better than the the military environment—and numerous studies original, and we repeat the process of reproduction especially motivated by military and civilian com- until we arrive at machines that satisfy our re- mand and control systems. It seems, however, that quirement. It has been shown that, in certain fairly possibly our knowledge of the human factor, spe- simple test cases, significantly improved perform- cifically in the man/machine environment, is less ance is obtained fairly rapidly. concrete than what we know about machines. These examples of procedures and concepts Let us consider an example taken, in simpli- that contribute to a seemingly intelligent behavior fied form, from a recent study. We consider this of machines, although far from exhaustive, may at situation: The subject is required to cope with least convey some idea of what machines can do an opponent who produces a series of events that and strongly hint at what they may be able to the subject sees as “data.” The subject is required do in the future. It is quite certain that their present to make certain judgments from these data, which main weakness—that they have to be told in too will presumably lead to some decision on his part. much detail what to do. instead of having the It turns out that human performance may be mark- capability to generalize, abstract, and adapt to edly superior or distressingly inferior to the per- unknown situations—will be continually improved. formance of a suitably programed computer, de- pending on the kind of judgment required. We will examine two simple cases. In both cases the opponent has a stack of playing cards C a p a b ilitie s o f tb e H u m a n from which he puts cards on the table for the sub- Partner ject to see (the “data”). For simplicity’s sake, we What, then, about the capabilities of the sec- concern ourselves only with black or red cards— ond member of the “team,” man? The study of no jokers, no values assigned to the cards. In the psychology, human-engineering, and behavioral first case, the number of red and black cards in THE SCIENCE FRONTIER 57

the stack is known to the subject. After each draw- of the best location of the interface between man ing the subject is required to estimate the prob- and machine. It is also quite obvious that the bur- ability of the outcome of that drawing. If the stack den of this determination is not so much on the contains, at the beginning, half black and half pure computer scientist as on behavioral scientists’ red cards, the estimate would be 50 per cent, re- understanding computers and on computer peo- gardless of the color of the card drawn. If. later, ple’s caring about human behavior and perform- the number of black cards drawn exceeds the num- ance. The results of such research are very likely ber of red cards drawn, and again a black card to have profound effects on future man/computer emerges, the estimate for the likelihood of that systems. Assuming, now, that we have determined event should be revised downward, and so forth. where man comes in and where the machine’s re- It turns out that in this game humans perform re- sponsibility, as it were, ends, there still remains markably well. the problem of making the two interact smoothly The second experiment is slightly different. at this interface. We have touched on this problem The opponent has several "strategies” (propor- before, but we shall devote the next section to a tions of black to red cards) at his disposal, from more detailed examination. which he can choose one for that particular game. Observing the outcome of successive drawings, the The Man/M achine Interface subject has to guess which strategy (distribution) the opponent is most likely to use in that game. There are two kinds of communication that Now, given that the probability of the opponent s have to be established between man as user, cus- use of each of his strategies is known, this can be tomer, and, hopefully, master and the machine computed from guessing games like those of the that aids him in his pursuits: (1) the program that first case. Humans perform quite poorly in this determines the structure, sequencing, and adminis- game, where the point really is. after each drawing, trative detail of the machine’s operation—possibly to assess the impact of the new event on the last at considerable levels of abstraction; and (2) the previous estimate. Machines do better, while hu- specific ad hoc mechanism of interaction that com- mans seem to be too conservative in changing their municates to the user specific results and the cur- opinions in the presence of a new event. rent state of the machine’s operation and allows This, to me, is a somewhat disquieting result. the user to convey specific instructions, data, and After all, a military commander is generally not operating commands to the machine. particularly interested in a numerical value for the likelihood of an observed enemy action but is programing languages interested in a fast and dependable assessment of the most likely strategy the enemy is pursuing. The We mentioned before that all internal infor- idea that perhaps a machine could do this better mation in the machine is couched in abstract than a human is disconcerting. Whether we like it machine code that has no direct relation to the or not, the line of demarcation between the realm problem the machine is working on. To run a of the human and that of the machine is by no specific problem on a machine, a scientist not means determined; we may be in for some consid- thoroughly familiar with computers will have to erable surprises. Remember, in this context, the use the intermediate services of an analyst and a clever and amusing book on strategies to win at programmer. The user is in effect separated by a blackjack that has, of late, attracted considerable great distance from the machine. To shorten this attention. Its strategies, developed with large distance, “problem-oriented languages” have been amounts of computer assistance, have caused gam- created whose object is to make it possible for ing-house operators to modify their rules, which the user to state his problem, or, more precisely, heretofore were perfectly adequate to ensure specify the desired procedure, to the machine in a profitable operation of the game with only human form as closely as possible related to the form in intuition, guesswork, and skill as opponents. which he would state the problem for his own use. It is. then, not surprising that present research A large number of such languages have come tries very hard to come to a better understanding into use, among the best known being fortran 58 AIR UNIVERSITY REVIEW

(FORmula TRANslation), alco l (ALGebraic languages is evident from this list of shortcomings. Oriented Language), and cobol (COmmon Busi- We would like the language to be more universally ness Oriented Language). Problem-oriented lan- applicable, thereby reducing the number of lan- guages are machine-independent, unlike assem- guages used, and closer to natural language. We blers, which are characteristic of their machine also need efficient means of translation from various type. Programs written in a problem-oriented languages into machine codes as well as from one language must be translated into the internal language to the other. In a way, this means trying language of the machine. This is done by compiler to have the cake and eat it. An easily translatable programs. Compiling occupies a significant part language tends to be extremely formal and, there- of the activities of a large-scale machine, often con- by, not too similar to human talk. Program lan- suming on the order of 30 per cent of the useful guages that come reasonably close to formalized machine time. Thus the economic significance of but intelligible English do so at the cost of being re- the compiling load is considerable. Nevertheless, stricted to a very narrow area of problems, as seen programing languages do facilitate our com- in our bridge language. Yet adequate solutions, munication with machines and indeed make the necessarily compromises, have to be found if com- programing of large-scale computing systems pos- puters are to play the roles of which they are sible. The advantages of procedure-oriented pro- capable in, say, military command and control graming languages are striking. First, they bring systems. the computer program into close relationship to This has led to intensive research in the sci- the problem as seen by the user; second, one in- ence of formal languages and their syntax and struction in the programing language text usually grammar. Such questions as equivalence of state- implies a large number of individual computer ments in different languages, translatability and instructions which are automatically detailed in relative efficiency, and effective ways of syntactic the compiling process; and, last, the use of a proper analysis are very much in the foreground of present programing language automates a number of tedious housekeeping tasks which must be done Figure 3. Com puter print-out in problem - explicitly when programing in machine code. oriented language for bidding at contract bridge For these advantages we pay a price. First, the advantage of close relation between the pro- bidrcl (/) = bidrcll (/; ptcount (ptsinsuits '/));* gram language and the problem implies that, to be distntuple (/); biddablesuits (/) efficient, the language is pretty much restricted to bidrcll (/; p; d; b) = Al, lessp (p; 12) preempt (/; ; p; d) the problem area for which it was designed. Thus Bl, lessp (nooflosers (/); U-5) greaterp (p; 2? arises the large number of languages that we have bid (1; CLUB) today and the need for numerous compilers and Cl, balanced (d; /)— i A2, lessp (p; IT) V greaterp (p; 20) — bid translators to make possible their use and the inter- (1; CLUB) change of programs. This process is cumbersome B2, T — bid (1; NOTRUMP) and costly. At the same time the languages that Dl, threesuithand (/)— A2, lessp (p; 17) — bid (2; CLUBS) we use to converse with machines and which they B2, lessp (p; 21)— bid (2; DIAMONDS) use to talk back to us are still far remote from the C2, T —‘ bid (1; CLUB) way we would like to talk, even if we agree to El, equal (length (b); 1)— „ A2. equal (firstsuit (b); CLUBS) V goodbutn accept considerable restrictions in our freedom of greathand (/;p;b) bid (l: lr3csuit (d))^ expression. Figure 3 shows a reprint of a program B2, T— bid (1; fir s ts u it (b)) intended to perform quite a human task, namely, FI, equal (secondsuit (b): CLUBS)— A2, greaterp (lengthsuitl (b) ; L) A greater*1 bidding at contract bridge. One sees that, although (lengthsuitl (b); lengthsuit2 (b)'i — there are numerous hints as to what is going on, A3, equal (firstsuit (b); DIAMONDS)— t this language has to be learned! In fact, some of (2; NOTRUMP) the major programing language manuals run to B3, T — bid (2; firs ts u it (b)) B2, T —bid (1; firstsuit (b)) hundreds of pages containing vast numbers of rules Gl, equal (lrshorter (b): DIAMONDS) A lessp (pj which have to be followed meticulously. bid (l; firstsu it (b)) The goal of present research in programing Hi, T— bid (1 ; lrs h o rte r (b)) research and promise significant results. One sees that many of these topics are just as applicable to natural language, and that, from the point of view of ideal communication between man and ma- chine, is as it should be. Language, after all. is the vehicle of our thoughts. The closer we come to making it flow freely from man to machine, the closer we come to the ideal of an integrated man machine team. language-oriented machines We noted that fundamental to our need for programing languages and associated translators is the fact that the machine converses internally in a purely machine-oriented language, the machine code. Perhaps our problem is that we try to make machines do things they were not really designed to do. Recalling that, basically, present-day ma- Figure 4. Direct communication with the com- chines have more than anything else the character- puter by cathode-ray tube display and light pen istics of a programed arithmetic calculator, one finds that their internal machine language could be the subject of continuing research covering almost considered as a close cousin to an arithmetic- all the aspects of computer and human behavior oriented programing language. It should, then, be we have discussed so far. possible to design a problem-oriented language of In computation centers, on the other hand, as much generality as possible (by incorporating where large-scale production of numerical results into it many of those characteristics that are com- is the goal, this interaction is quite restricted. It is, mon to many of the existing more specialized however, recognized that, even in purely computa- languages) and then to design a machine that uses tional tasks, the close ad hoc interaction of man this language as its internal means of communica- and machine can be of extreme value. (Note the tion. Obviously such a machine w'ould not need a example of Figure 1.) For instance, machine pro- compiler, since external and internal language are grams exist which, given a sketch of some mechan- identical. ical structure, say, a girder, and the load on it, will At least one such language and its machine produce all the information for the exact construc- implementation have been designed, and the ma- tion drawing, including dimensions, stresses on all chine is presently under construction. Simulation the members and their connection points, and so experiments have shown its significant superiority forth. Logically, it would be desirable to display over conventional machines, particularly in non- this computer-generated drawing to the designer numerical processing. as a drawing, not as a sheet full of printed figures, and to give him a means to sketch in changes and observe the result. This is indeed being done, the direct communication means being Tv-like cathode-ray tube displays and Direct communication between user and ma- “light pens” allowing the user to introduce data chine usually occurs by means of keyboards, suit- and instructions into the machine by drawing able displays, and printing devices turning out graphs on the face of the tube. The same procedure results in the form of printed records. In military proves of great utility in mathematical computa- command and control systems, where real-time tions of complex situations where unfolding proc- interaction between humans and machines is of esses can be observed graphically and controlled vital importance, the means of intercommunica- by the operator. Figure 4 illustrates such an ar- tion are frequently extremely elaborate. They are rangement. 60 AIR UNIVERSITY REVIEW

In many applications, this and related tech- I his leads to the concept of all sorts of parallel niques of linking man and machine are attracting processors which in their practical realization run increasing attention as one of the most promising the gamut from relatively straightforward con- ways of teaming the capabilities of the human and figurations of multiple arithmetic units to assem- the automaton. blies of full-size computers that can operate inde- There are limitations to this direct coopera- pendently. These machines converse with each tion. Technically, there is the disparity of operating other and merge their operations where appropri- speeds, hi the time scale of computer operations, ate by way of common memories and under control the time an operator takes to make up his mind of a central unit that can be visualized as a com- about what to do next is a costly eternity. The bination of supervisor and traffic cop. monetary value of a minute’s running time of a In other cases the restrictions imposed by the large-scale computer is about $10. This suggests, rigid organization of existing computers are in the for economy’s sake, the use of relatively small com- center of attention. Machines are contemplated puters or time-sharing of larger installations, where that are capable, to a certain extent, of varying during waiting times the computer does something their internal organization as required by the prob- else, such as running a production program or lem at hand. The language-oriented computer con- serving other customers. Meanwhile the individual cept described earlier is another instance of the user retains the impression that he is in control of strong interaction between the requirements of the the machine. Several time-sharing systems of this user and the machine organization. The develop- kind are in experimental operation. In systems with ment of content-addressable memories, i.e., memo- many sharers, things can become a little compli- ries that can be queried by the content of their cated. Not only must the machine do the work, it entries rather than an address location, are an- also has to allocate time to users, minimize waiting other example of the user’s demand triggering new times, and—last but not least—make out the bills for developments. Content-addressed memories have services rendered. Thus a sizable part of its ca- important potential application in the field of pacity may be consumed by administrative chores. recognition systems, in data storage and retrieval, and automatic language translation. Frequently, of course, such functional changes demanded by new uses are difficult or impossible to implement by New Concepts—New M achines conventional electronic techniques. The content- addressed memory is an example. Sometimes the the influence of new applications interaction also occurs in the opposite direction. In our discussion so far, we may have—unin- New and promising physical technologies emerge tentionally—created the impression that our current that demand unconventional machine design in computers are more or less set in concrete and order to be used to their best advantage. In the that all we can do is to make the best possible use next section we shall look at an example. of them the way they are. Nothing can be further from the truth. Two interacting factors continually the influence of new techniques force new developments in the internal concept as well as the physical embodiment of new machines. It may have seemed strange to some readers One is that new machine applications demand that we have been discussing computer sciences properties of our machines that are not present in for so many pages without paying too much atten- conventional computers. For instance, we notice tion to the hardware aspects of computers. The that in many applications already discussed we reason is that at the moment most of the really have to perform fairly simple operations on very burning questions are in the area of proper use and large amounts of data. Very frequently, as in the organization of computers rather than in their case of pattern-recognition schemes, these opera- implementation. Yet in electronic techniques sig- tions are individually independent of each other nificant changes are taking place that cannot fail and could very well be performed simultaneously. to have profound effects on the design of future This is also true for many computational tasks. machines.

Figure 6. Model of an iterated net using conventional components

Figure 7. The black rectangle on the wafer is the microcircuit equivalent of one segment of the net in Figure 6. The ceramic holder, for easier testing and handling, is eliminated in actual use. THE SCIENCE FRONTIER 63

Various new techniques are either available ing mode. This seems to indicate that these tech- or in the development stage: integrated circuits, niques may be particularly suited for applications cryogenic devices using the properties of materials where spatial relations or structured data fields are at extremely low temperatures, and devices using involved, and this is the case in many adaptive and processes in thin magnetic films for information- pattern-recognition systems and in instances where processing operations. All these techniques, of some sort of field representation is indicated. which microcircuitry is advancing quickly in the So far, only modest models have been built application field, have in common that individual and tested in the laboratory. Figure 6 shows a net devices are extremely small in size and are pro- of this kind, built of conventional components and duced in bulk processes as opposed to wiring containing only 16 elements. If light is projected operations connecting lumped components as in through a properly punched ib m card (not shown conventional circuitry. By use of this technique, a in picture), the net can be made to act as a counter single transistor case the size of a pencil eraser or an oscillator, or a variety of other computer can replace an entire plug-in module of the con- modules. A single element, embedded in a ceramic ventional variety. Without packaging, the picture holder for easier testing and handling, is shown in is even more impressive. Figure 5 shows a silicon Figure 7. wafer containing about 50 integrated circuits, each An interesting and important problem in this of which is capable of replacing the plug-in module connection is that of reliability. If we make it our shown alongside. Further, in the unpackaged form goal to produce large numbers of logic elements these circuits, in quantity production, promise to on one common substrate, we cannot possibly ex- be extremely cheap and very reliable. Properly pect to have no faulty elements at all. For reasons used, they seem for the first time to give us a means of economy, it will not be feasible to discard entire of producing entirely different computers without arrays only because of a feu faulty elements. Even too much restriction on the number of com- in the presence of very good yields in terms of ponents. individual elements, the discards would outnumber “Properly used” here means that, first, we use the good arrays astronomically. Also, it might be elementary circuits which permit us to build up extremely difficult to conduct exhaustive tests to any computer structure using one kind of element locate faulty elements. It is therefore necessary to only. Such elements do exist. Second, large num- design structures that continue to function reliably bers of such elements should be manufactured on in spite of failures of a number of components. a common substrate with all conductors that might Since we have excluded outright repair of possibly be needed to interconnect elements on the faulty elements, two approaches to our goal re- wafer produced in the manufacturing process. Wir- main: self-repair and fault-masking. In the latter ing in the traditional sense would be used only at case, we leave the faulty elements in the structure the periphery of an entire array of elements. By (as a matter of fact we have no means of telling additional peripheral coded inputs or by signals that there are faulty components), but we have a requiring no physical contact, such as light patterns sufficient number of additional, redundant elements or local magnetic fields, the array can then be to maintain the correct function of the array. One “structured,” i.e., instructed as to which connec- way to achieve this, which is particularly suited for tions should be active and which should be inca- very large nets, has recently been under investiga- pacitated. tion at afcrl. It is, in essence, a generalization of Logic structures of this general kind have the technique applied to relay nets by Shannon and been investigated quite extensively in the recent Moore in 1956 and applies the principle of iteration past. We now have, for the first time, promising as shown in Figure 8. One notes that in 8a three techniques to put them into actual operation. We elements, shown as circles, are connected in a cer- can visualize that in structures of this kind many tain way. In 8b we have three configurations different processes can take place simultaneously identical to 8a, which in turn are connected to each in different regions of the array and, further, that other in the same way as the elements were in the processes involving interactions of happenings in original net. In other words, 8b results from 8a by neighborhoods are most likely a preferred operat- the replacement of each element by the net itself; c results from b in the same manner. In this way and amplifying elements to a circuit. In many cases arbitrary large nets can be “grown” and their be- such lumped components are no longer identifi- havior studied. It turns out that by use of this able; the electronic action takes place in the dif- technique arbitrarily reliable nets can be construct- ferent regions of a block of material which by de- ed from arbitrarily unreliable elements. Research signer’s choice have been given certain geometries is proceeding to make practically significant uses and different electrical characteristics. It becomes of these techniques. more appropriate to describe the action of the de- vices in terms of the basic physics involved, e.g., in terms of diffusion and combination processing. S pa c e does not permit me to relate The voltages at the output terminals—if output a large number of equally significant new elec- terminals there be—result from associated field dis- tronic techniques and their impact on future com- tributions. Thus physicists take over, at least in puter system design. For the same reason I was part, functions that previously were considered the forced to omit a host of important computer appli- domain of circuit designers. cations, each of which is justly deserving of sepa- Let us turn next to the maker of computer rate treatment. The areas of information storage languages. It might appear on first sight that the and retrieval and automatic translation of lan- specification of a programing language is a matter guages are but two instances. of concern just to an experienced programmer who In conclusion, let us only briefly consider an knows what he wants and has an adequate appre- aspect of this research towards better and more ciation of the intended field of application of the useful information-processing systems, namely, the language. What he is doing, however, has wide- interdisciplinary character of this research. Obvi- spread ramifications touching on widely separated ously, electronic engineers and circuit designers, as branches of science. First, he is indeed constructing well as logic designers mostly with engineering or a language, a system of nouns, verbs, and modifiers, mathematics background, have had and will con- governed by grammar and syntax. In fact, since the tinue to have a deciding influence in the physical goal of the programing language is to close the gap design of computers. But consider the advent of between natural and artificial languages, a con- integrated circuitry! As bulk processes of fabrica- tinuing study of natural as well as artificial lan- tion take over, it will become increasingly mean- guages is necessary. Also, since artificial languages ingless to design “circuits” in the conventional are governed by rigid rules allowing no exceptions, sense of combining resistors, capacitors, inductors, they present a proving ground for a branch of sci- THE SCIENCE FRONTIER 65

ence called mathematical linguistics concerned line between human intellect and that of a machine with formal methods and algorithms of syntactic is delineated by the ability of humans to devise analysis. This, in turn, has important implications such nonprovable theorems and the lack of that for the analysis and translation of natural lan- ability in machines. guages, one of the more ambitious goals ol com- From their very beginning, computers have puter applications. attracted the interest of psychologists. The emer- On the other hand, programing languages, gence of controllable machinery that displays such being systems for manipulation of symbols and behavioral traits as “learning” and conditioned re- groups of symbols according to a set of fixed rules sponses and that is able to perform in a primitive and based on certain axioms, are, along with the way such feats as abstraction and concept forma- internal arithmetic of a machine, formal systems in tion suggested computers as welcome tools for the the mathematical sense. This exposes them, in modeling of certain ways of behavior. (“Assuming principle at least, to the problems and questions that a rat traversing a maze reasons in a certain associated with formal systems, which have rocked way, how would it perform?” A computer simula- the world of fundamental mathematics since the tion can give a quick answer.) More important are start of the century. For instance, a programing the questions of relative effectiveness of humans language must, first of all, be consistent. Its rules and machines. More research in that area is of ut- must not permit statements contradicting each most importance. Further, what are the uses of other. Further, people concerned with the funda- computers in the field of teaching and training? mentals of mathematics are very much concerned (To mention this topic this late and so briefly is with the question of computability, i.e., whether almost inexcusable in the light of the fact that a certain function can be computed within the major manufacturers in the computer field expect framework of a given formal system. The relation a very substantial portion of their future business to computer sciences is obvious. In fact, Turing’s to be in the area of mechanized education. Thus hypothetical machine, mentioned at the beginning it is not too surprising that a considerable number of this article, was not conceived at all as a design of researchers in the computer field have a back- for a computer but as a means to axiomatize the ground in psychology or behavioral science.) problem of computability. Lastly, let us mention the intimate relation Let us revert for a moment to the question between computer sciences and the life sciences. of consistency of formal systems. It has been Early catch phrases about computers as “electronic proved in Goedel’s famous proof (1932) that a brains” outthinking humans have, fortunately, formal system must contain at least one statement largely disappeared. But it takes not too much which, although true, cannot be proved to be true imagination to see the analogy, in performance, by means of the system itself. In fact each non- not in physical construction, between the large trivial® system contains a large number of such logic nets discussed previously and a nervous struc- statements. There exists a sizable number of ture in a living animal. Both have the capacity to theorems in mathematics which have not been dis- perform certain logic or cognitive functions and proved but for which a formal proof has not been to continue to do so in spite of local malfunction. found in spite of continuing efforts of the most Many concepts in advanced computer organi- able mathematicians. One fundamental law con- zation have taken clues from the life sciences. In cerning formal systems tells us that indeed an in- one field, known as bionics, a search is being con- finity of such propositions does exist in any formal ducted for biological mechanisms performing cer- system rich enough to describe a Turing machine, tain data-processing tasks that can be duplicated which, in turn, can be taken as a representative in electronic hardware. Despite the objection that abstraction of any existing computer. These sup- may be raised that, after all, the biological way is posed theorems are not provable by the mechanism only one way to do things and not necessarily the of the machine by virtue of the fact that such best way, biological systems are nevertheless a “theorems” are “undecidable.” As has been pointed valuable way to gain inspiration. Less teleologically out by Cannonito, this may suggest that the border oriented research in biological systems serves to °A system rich enough to define, say, arithmetic in it. verify basic hypotheses of information sciences by 66 AIR UNIVERSITY REVIEW investigation of the performance of the living of the direction in which I believe computer re- system. Hence our intense interest in the sensory search will go: systems of animals and humans. On the other hand • There is a great deal to be done in im- the techniques of modern information science pro- provement of physical techniques. Some of these vide unprecedented tools to the researcher in the improvements will increase the capabilities of con- life sciences. This is increasingly being recognized. temporary machines. To a large extent improve- To prove that this is not a unilateral feeling, I quote ments will take place under the pressure of compe- from the anatomist D. A. Sholl’s well-known book, tition and the need for more capacity along estab- The Organization of the Cerebral Cortex: lished lines. As a final example of the application of • New techniques make possible, and may recent mathematical developments outside even demand, differently structured machines. their original domain, we shall note the investi- This, together with the shift from numerical proc- gations of the relationship between logic and essing to the accomplishment of tasks closer to mathematics. Such a study, one of pure mathe- human reasoning and decision-making, may lead matics, might seem to be especially remote from to an entirely new generation of machines. the experiments of the biologist. But the theories of language and logic that emerged • The relation between man and machine from this work have had a far-reaching influ- is in the process of being radically revised. The ence. Apart from the purely philosophical move towards a free flow of language between studies, the influence of this work extends from user and machine is probably the most important the symbolic expression of biological theories topic in present research. by Woodger to the development of high-speed • The need to draw on the resources of a digital computers. considerable variety of branches of science en- courages the re-emergence of scientists with an M any topics worthy of detailed discussion in their interdisciplinary outlook. It may even encourage own right have not been mentioned at all; others the proponents of specialized branches of research have been treated in perhaps undeserved detail. to mend their ways and try to speak in more uni- The reasons for both these shortcomings are, pri- versally understandable language. Although a re- marily, the author’s personal leanings and the turn to the universitas litterarum is impossible, desire to present at least some items in detail rather research in computer sciences may prove to be a than to attempt to give an exhaustive list. Be that considerable force towards cooperation and mutual as it may, I feel that I owe the reader a summary understanding among many branches of science. Air Force Cambridge Research Laboratories ARE WE MISSING THE BOAT IN MANAGEMENT DEVELOPMENT?

M ajor Robert L. A bl e

VERYWHERE in the literature of today’s cess in other areas are nullified. It is through peo- E Air Force one finds repeated references to ple that we manage and accomplish our required the great requirement that now exists for sound tasks.” Citing the steady escalation in the education management practices in the Air Force. General levels (and requirements) of Air Force people, the l^eMay, in a recent speech before the American Chief of Staff went on to stress the demand which Bankers Association, set this requirement into exists for continuous training: “We in the Air perspective by citing as criteria the capital invest- Force are taking advantage of every means avail- ment of the Air Force, the number and dispersal able to increase the effectiveness of our manage- of its physical facilities, the size and versatility of ment capability at all levels.” It is the purpose of personnel strength, the size of the annual expendi- this paper to suggest that Air Force supervisors in ture budget, and the dynamic nature of our mission the field are not utilizing some of the most effective which requires us to stay always ahead of our com- and available tools of management development, petitors in the development of the tools and tech- even though encouraged by the Chief of Staff and niques of modem warfare. If we accept these cri- in some cases required to do so by Air Force teria as evidence of the requirement for sound directives. management, it is immediately apparent that the To ensure a continuously available supply of Air Force qualifies on all counts to a greater degree adequately prepared people necessitates a planned, than any private corporation in the United States. overt concern for their development. At the present In the same speech. General LeMay reminded time the Air Force does not have a specific, inte- his audience that our most important resource is grated program of management development. people. He pointed out that “if we fail to use our However, it is not the “program” aspects of such human resources properly, all prospects for suc- an endeavor which produce the desired results. 68 AIR UNIVERSITY REVIEW

Effective management development results from rational decision-making. Just as an effective path the actual practice of the philosophy and methods up the ladder in those days was the study and of developing the executive talents of management emulation of the successful leader, so today is there personnel. In a sense, management development is a heavy requirement on the aspiring subordinate a climate, an environment in which commanders to learn the skills of his manager superior. accept as an integral part of their responsibilities Internal promotion necessitates internal de- the development of their subordinates’ abilities. velopment. A significant case can be made for The command line is not only charged with the management development on the grounds that the responsibility of building an effective organization alternatives are unsatisfactory. The requirement but also with maintaining and perpetuating that for the managerial skills is self-evident for any or- level of effectiveness. Thus, to fail in the organiza- ganization. For private firms, the principal alterna- tion’s perpetuation because of a lack of well- tive to internal development is to hire away the trained and developed people is as clearly a com- necessary skills from other firms. The steeply mand failure as would be the failure to maintain progressive income tax makes this a very expensive the organization’s equipment or physical facilities. alternative. Furthermore, an expanded use of stock A good commander would never let his buildings options, profit-sharing retirement programs, and deteriorate through neglect, to save the expense similar postwar innovations designed to reduce of preventive maintenance. Similarly, a good com- executive turnover makes this alternative even mander should never let his people deteriorate more unsatisfactory. For the Air Force, the policy through neglect, to save the time necessary for of exclusive promotion from within leaves no al- their development. It is the purpose of this paper ternative. We either develop the skills or do with- to measure the extent to which Air Force super- out them, and the stakes have become too high visors are actually using the methods of manage- for the latter course. ment development available to them. Ceneral Electric conducted an intensive study of the paths of ascension followed by its 300 top significance of management development managers. A principal conclusion of that study was that the natural selection process (or do nothing The significance of management development and hope the cream will rise to the top) supplies in private corporations can be amply documented. neither the number nor the competence level in Evidence that the demand for managerial talent even a minimum adequate quantity.4 The Air is increasing more rapidly than for the labor force Force has clearly rejected the “natural selection in general is reflected by the fact that managerial process” as an adequate solution to the problem jobs increased 89 per cent from 1940 to 1963 while and recognizes that there are few “born managers’ ; the labor force increased but 32 per cent and the most are developed.5 general population increased 43 per cent.1 An The decade of the 1960’s presents a particu- American Management Association study revealed larly critical personnel problem for the Air Force overt management development activity at 78 per because of the very large proportion of its officers cent of the companies with over 10,000 employees.- who entered active duty during World War II. The National Industrial Conference Board con- These officers have progressed into key positions ducted an international study in eight nations and of command and management and thus embody a reported a consistent and pervasive concern for substantial proportion of the executive experience the problem.3 Thus industry has recognized the of the officer corps. This cluster of experienced changes which have taken place in the composition officers, usually referred to as the “hump problem,” of its work force and has responded with an in- is reflected in the number of officers 40 yeais of creased emphasis on the development of mana- age or over. The percentage of Air Force officers gerial talent. A similar composition change is tak- 40 or over, as of 31 May of each year, is as follows: ing place within the Air Force as the World War 1956 1957 1958 1959 1960 1961 II blood-and-guts form of charismatic leadership is replaced by the manager skilled in the tools of 13% 17% 18% 24% 29% 31% IN MY OPINION 69

The real significance of these figures lies in selection of these methods. For one thing, current their relationship to the Air Force retirement sys- Air Force directives either recommend or direct tem rather than to any diminishing capability of their use by Air Force supervisors. Also, none of officers over 40 years of age. This “hump" group the selected methods are restricted by direct budg- began reaching 20 years’ active service in 1960 etary limitations; therefore, the Air Force has and will therefore be reaching the customary re- within its own command prerogatives the means tirement period during the decade of the 1960’s. for stimulation and encouragement of a greater This condition is reflected in a recent survey of use of these methods. Finally, since all of the 11.499 Air Force officers, which reports the per- selected methods are accomplished on-the-job centage of each rank planning to retire before (with the partial exception of self-development), 1970:« an increase in their utilization involves no new Lt. All Officers programs or directives but only increased support Colonel Colonel Major (except Generals) by all levels of command. 72% 73% 74% 40% Before discussing the results of this study, we might well establish a common ground of While similar figures are not available for general meaning for each of the selected methods. The officers, their rate of retirement during this period, definitions finally accepted for these methods are because of even longer service, would be expected the product of an extensive research of the litera- to be similar or greater. Thus, during the current ture of management and are usually consolida- decade, the Air Force must replace nearly three- tions of many different approaches. fourths of its key officers. So, also, must the ex- perience level lost with these officers be replaced Counseling. Counseling seeks to change be- if a deterioration in effectiveness is to be avoided. havior, attitude, or point of view; focuses on ob- jectives, goals, hopes, fears; is person-centered, use of selected methods of management not job-centered. This type of counseling is con- development ducted by management personnel in the normal course of their responsibilities for getting things With all this support and urgency for manage- done through people. Thus it should not be con- ment development, one would certainly assume fused with counseling conducted by. clinical psy- that the Air Force was actively engaged in the chologists or psychotherapists who seek to obtain various processes that help achieve the stated deep personal adjustment. objective. A review of Air Force directives tends Coaching. Coaching has been variously de- to support General LeMay’s assertion that we are fined but always with major emphasis upon the taking advantage of even.’ means available to in- interrelationship which exists between the supe- crease the effectiveness of our management capa- rior, the subordinate, and the job. Coaching is bilities. The real question which must be answered, here defined as planned, on-the-job experience to however, is to what degree “every means avail- impart skills, teach a job, convey facts; focuses on able" is actually being utilized by Air Force super- methods; is job-centered, not person-centered. visors to aid in the development of their subordi- nates’ managerial skills. Understudy. The most frequent conception To find the answer to that question, the of understudy training postulates the assumption author selected five of the methods most frequently of the superior’s role by the subordinate. For the employed by industry to develop the managerial purpose of this study, the Air Force definition is skills of its executives. A study was then conducted used, viz., training to assume the responsibilities to determine the extent of their utilization by Air of the supervisor, either as a substitute or as a Force supervisors. The five methods selected were replacement. counseling, coaching, understudy training, job ro- Job Rotation. There can be little question tation, and self-development. In addition to being regarding the value of learning on the job as a the methods most frequently used in industry, a principal means of developing the managerial number of other factors served as the basis for talents of executives at all levels of an organiza- 70 AIR UNIVERSITY REVIEW tion. Job rotation is one of the methods of manage- is his own responsibility, and the Air Force has no ment development which incorporates this prin- intention of "spoon feeding” him. While the organ- ciple of learning-by-doing in such a way as to pro- ization should provide guidance and assistance in vide experience in a variety of jobs. Some of the developing a man’s abilities, the man himself must conceptions of job rotation found in the literature activ ely participate in the process of preparing for specify that the movements be planned. However, an ever increasing level of responsibility.7 from the standpoint of developing the managerial There are many methods available to the in- abilities of the members of an organization, the dividual for developing his managerial abilities. really crucial point is that the member receives However, this study is restricted to those tangible the broadening experience. It is thus most im- activities which can be measured through depth portant that the jobs carry full responsibility, that interviews by an interviewer untrained in psycho- the responsibility is progressively increased with logical investigation. These activities include off- each job change, and that the jobs are sufficiently duty college courses, professional reading, profes- differentiated as to require the exercise of different sional societies, and community organizations. abilities. If these conditions are met, the move- ment of personnel through job rotation becomes a tool of management development, whether or management development activities not such movement is planned solely for that pur- of Air Force supervisors pose. Since the Air Force does not conduct a pro- gram of job rotation planned solely and specifically To determine what is actually being done in for the purpose of personnel development, the the Air Force with reference to the five methods "planned" portion of these definitions was not of management development under study, we carried forward to the composite definition of job sought a sample of experience of Air Force officers. rotation used in this study. Here job rotation is A sample was desired which would satisfy five defined as the provision of broad experience basic criteria. First. The most important character- through periodic movement fo increasingly re- istic desired in the sample was that it be sufficiently sponsible jobs requiring different abilities. dispersed so that no single command policy, com- Self-development. Throughout the literature mander, or supervisor would exert a disproportion- of management, one continually finds expressed ate influence on the results. Second. A distribution the concept that management development is in- between flying and nonflying officers closely paral- dividual development. The final responsibility for leling the total Air Force distribution was desired. development lies with the individual himself rather The purpose of this criterion was to prevent the than with the organization. This is not to say that the peculiarities of a few assignments from distorting organization carries no responsibility for the devel- the results. Third. For the same reason, a distribu- opment of the abilities and potential of its people. tion similar to that of the total Air Force was The organization can create the proper climate by sought between operational commands (those encouraging developmental activities, providing charged with a combat mission) and support com- for advanced education, and generally supporting mands (those charged with a noncombat mission). the program. For example, the counseling, coach- Fourth. A sample was desired which reflected the ing, understudy training, and job rotation previ- total Air Force distribution for rank. Fifth. Finally, ously discussed are very valuable media through a sample of approximately 100 officers was con- which the individual’s self-development can mate- sidered large enough to report reliable answers rialize. Yet in the final analysis the individual con- but small enough to make possible the intensive, cerned must exert significant efforts if any real face-to-face interviews considered necessary to ob- degree of success is to be achieved. The Air Force tain valid answers to the questions posed. recognizes this dual responsibility of the organiza- An organization which closely conforms to tion and the individual by providing the oppor- these desired criteria and chosen for the sample tunities for development and then leaving it to the in this study was the student body of the School individual to make use of them. An officer’s career of Engineering at the Air Force Institute of Tech- IN MY OPINION 71

nology ( afit), Air University Command, YVright- as “very typical” of their careers and not the result Patterson Air Force Base, Ohio.® All five of the of an unusual circumstance during the year re- criteria were amply satisfied by the sample chosen. flected by the interview. Some 68 per cent had A significant advantage of this sample is the not been kept continually informed regarding their fact that these officers represent a group which strengths, weaknesses, and level of achievement, the Air Force has identified as having definite their supervisors disregarding the “good manage- potential for major contribution to the organiza- ment practices” recommended by the Air Force.8 tion in the future. This is evidenced by the highly Furthermore, 80 per cent of the respondents had selective process used in approving afit students not been counseled and assisted by their super- and by the fact that the Air Force is willing to visors in developing long-range career plans, and spend a considerable amount of money on their only 27 per cent had been encouraged to make advanced education. It logically follows that if the specific career progress plans. The great bulk of Air Force is putting into practice the methods of supervisors had thus been disregarding the Air management development described in its direc- Force’s recommendation that “all supervisors coun- tives, it should certainly be applying them to its sel and assist subordinate officers in developing relatively junior officers with recognized potential. long range career plans.”9 The 28 February 196J A face-to-face, personal interview was held edition of the Air Force sample survey asked with each officer included in the sample. These 11,934 officers if their supervisors had “counseled interviews were directed discussions in the sense or assisted you in developing long range career that a questionnaire was used as an interview out- plans.’ 10 Some 87 per cent reported that they had line. They were nondirected discussions in the not received such counseling, which compares sense that the respondents were encouraged to talk favorably with the 80 per cent negative response at length and volunteer whatever information they reported in the present study. Several other studies desired.00 An analysis of the results of these depth report similar results.11 interviews revealed a general failure of supervisors Further analysis of the data revealed that 94 to use the selected methods of management devel- per cent of the respondents felt that such counsel- opment, w ith the exception of understudy training. ing would have been very beneficial to them. Use of Counseling. Counseling has not been Analysis also revealed that when Air Force super- used by most Air Force supervisors as a method of visors do attempt counseling they do a good job of developing the managerial skills of their subordi- it, as indicated by the opinions of their eounselees nates. While some 64 per cent of the respondents’ and by the degree of conformance to the perform- supervisors had made some attempt at discussing ance standards. For clarification, it should be noted their annual evaluations with subordinates, 80 per that the respondents of this sample were reporting cent had made no attempt to use the discussion as on a period in which the Officer Effectiveness Re- an opportunity to counsel the subordinate regard- port ( oer) was shown to the ratee. However, if ing his strengths, weaknesses, and performance. 80 per cent of the supervisors were not counseling The respondents reported this failure to counsel when such a session was prescribed, on what basis could it be argued that they were doing so six •It should be clearly understood that the officers comprising months after the oer, as now stipulated in Air this sample were interviewed regarding experiences which had occurred prior to their coming to the Institute. Therefore, the Force Manual 36-10?1- results of this study in no way reflect the practices of AFIT. the Air University Command, or Wright-Patterson Air Force Base, Use of Coaching. Coaching has been simi- beyond the contribution made by any former member of these commands who happened to be included in the sample. larly neglected. Some 56 per cent of the respon- dents had not received any kind of coaching, and * #Fnr those interested in the research method used. .» brief outline of the format is given. An interview guide questionnaire 78 per cent had not received coaching designed to was constructed to ensure a common ground of meaning for all teirninology and to ensure that the same appropriate questions, develop their managerial skills (the difference be- using the same wording, were asked all respondents. Composite conceptions of how each of the selected methods should be ac- ing the coaching received in aircraft checkouts). complished were consolidated from the literature of manage- Since this was “very typical” of their careers, the ment. These composites served as the performance standards against which the respondent's experience was compared. An great bulk of the officers had not received the affirmative response was not recorded unless the performance standards had been reasonably satisfied. “assistance in developing and understanding the 72 AIR UNIVERSITY REVIEW skills required for managing both the human and sued a planned program of professional reading, procedural aspects of operations,” as recommended but only 6 per cent had been encouraged to do so. by the Air Force13 and as strongly endorsed by the officers themselves. The respondents’ opinions and the conformance to the performance standards S ever a l im po r t a n t points made here should be again reflected a good job of coaching when at- restated for emphasis. For one thing, the Air Force, tempted. as a very large organization charged with a vital and dynamic mission and responsible for a stagger- Understudy Training. Air Force supervisors ing capital investment and a substantial annual generally do prepare an understudy to take over expenditure, has a very real and continuing re- their responsibilities, either as a substitute or as a quirement for a large body of officer-managers replacement. Some 42 per cent of the respondents possessing a high level of proficiency in managerial had been so prepared, and most found this “very skills. In order to get such officers, our policy of typical” of their careers. While it is true that the promotion from within leaves us no alternative but majority (58 per cent) had not received the train- to develop them ourselves. Toward that end, Air ing, it should be noted that this training is designed Force directives clearly prescribe the philosophy to be given to one subordinate by each supervisor, and some of the activities which private industry rather than to all subordinates as is true for coun- has found to be successful. Unfortunately, no seling and coaching. For example, the Air Force philosophy however conceived, no endorsements states that "any supervisor should provide someone however sincere, and no activities however effective to act in his place.”14 Thus it is important that the will ever elevate the skills of a single manager un- training be provided by each supervisor, rather less given meaning by the actions of supervisors in than be received by each officer. The degree of the field. The results of the study reported here, conformance to the performance standards devel- as well as other supporting studies, clearly reflect oped in the literature of management indicated a general lack of such activity on the part of Air that a good job of understudy training had been Force supervisors. done. This neglect would be the more excusable Use of Job Rotation. In the opinion of 56 per were it not so easily remedied. We could do the cent of the respondents, the Air Force had not job if Air Force supervisors would realize that the rotated their assignments to develop their mana- perpetuation of organizational effectiveness is as gerial abilities. Furthermore, 59 per cent had not dependent upon the development of human skills received the variety of experience which the Air as it is upon the development of weapons. This Force deems appropriate for their rank.15 realization would go a long way toward creating Interest in Self-Development. Analysis of this an environment of development. In such a climate, sample revealed that Air Force officers have been numerous opportunities would appear for coach- far more active in trying to develop themselves ing; prescribed counseling sessions would become than their supervisors have been in encouraging meaningful; encouragement would sow the fertile them to do so. Some 53 per cent had not received field of self-development; understudies would be any encouragement to pursue any of the self- prepared to assume the superior’s responsibilities; development media under consideration. While 61 and a variety of experience could be provided that per cent had taken usafi or ec i correspondence would form the foundation of a career progression courses during the previous year, only 26 per cent substantially escalated from the parochial paths had been encouraged to do so. All of the respon- followed by so many today. dents had applied for afit, of course, but only 35 What is needed is a pervasive awareness of per cent had been encouraged to do so. While 32 the need and importance of developing the man- per cent had taken off-duty college courses (some agerial skills of all officers in the Air Force. To had been stationed in remote areas where no such obtain such an awareness, we need the active sup- opportunities existed), only 22 per cent had been port of all commanders at all levels of the Air encouraged to do so. Some 25 per cent had pur- Force. Success in such endeavors lies easily within IN MY OPINION 73

our ability and within the directives and programs the Air Force to give real meaning to the words we now incorporate. It is up to the supervisors of of their Chief of Staff. United States Air Farce Academy

Notes 1. U.S. Department of Commerce, Statistical Abstract of 8. AFM 36-10. Officer Effectiveness and Training Reports the United States ( 84th annual ed.; Washington: U.S. Govern- (Washington: Department of the Air Force, 25 August 1960), ment Printing Office, 1963), pp. 5, 219, 232. p.12. 2. Joseph M. Trickett, A Surrey of Management Develop- 9. Air Force Regulation 36-23, “Officer Career Manage- ment: The Quantitative Aspects (New York: American Manage- ment,” 8 May 1963, p. 2. ment Association, 1954). p. 24. 10. The USAF Personnel Report, VI, 2 (1 September 3. Walter S. Wikstrom, Developing Better Managers: An 1961), 20. Eight-Nation Study (New Y'ork: National Industrial Conference 11. Leighton F . Downing, “Executive Development in the Board. 1961), p. 3. United States Air Force" (unpublished thesis, Air War College, 4. Harold F. Smiddv, “General Electric’s Philosophy and Air University, Maxwell Air Force Base, Alabama. 1955), pp. Approach for Manager Development,” Fitting Management De- 46, 47; Melvin Max Burnett, "Analysis of Career Development velopment to Company Needs (General Management Series No. Opportunities for Personnel Managers in the United States Air 1T4; New York: American Management Association, 1955), Force” ( unpublished master's thesis. School of Government, p. 14. George Washington University, 1958), p. 110. 5. For example see Air Force Manual 25-1, The Manage- 12. AFM 36-10 (25 July 1962), p. 6. ment Process (Washington: Department of the Air Force, 13. AFM 25-1, p. 82. September 1954), pp. 79, 85. 14. Air Force Pamphlet 50-2-24, “How To Develop Un- 6. The U.S. Air Force Personnel Report, VII, 3 (Washing- derstudies and a Plan for Self-Development" (Washington: ton: Department of the Air Force. 1 Februarv 1962), 34. Department of the Air Force, 1 June 1955), p. iii. 7. AFM 25-1, p. 84. 15. AFR 36-23, pp. 2, 3. Air Force Review

POSTGRADUATE OFFICER TRAINING FOR PILOT TRAINEES

C aptain A lan J. Grill

F F IC E R training in Air Training Com- Air Training Command. This apparent overlap of O mand’s undergraduate pilot training pro- instruction, combined with the fact that approxi- gram has recently undergone an extensive revision mately 90 per cent of students entering pilot train- —a revision necessitated by changes in the currieu- ing today have college degrees and appear to be lums of the precommission schools and by a grad- adequately schooled in the basics of communi- ual increase in the educational level of the student cative skills, caused Air Training Command to pilot input. In revising the program, sound reason- direct an overall comparative study of officer train- ing and research led to the formulation of a “post- ing being conducted in the precommissioning graduate” philosophy for the officer training por- schools and in upt. tion of the undergraduate pilot training ( upt) In the course of this study two conflicting program. concepts appeared. One school of thought advo- The precommission schools that constitute the cated the theory that, since present student input primary sources of student input for the upt consists exclusively of commissioned officers, no program are the Officer Training School ( ots), further officer training other than general military the Air Force Reserve Officers Training Corps training is necessary in upt. The opposing school of thought contended that the officer training con- ( afrotc), and the Air Force Academy. Studies of their eurriculums revealed that the precommis- ducted by the precommission schools was broad sioning schools were providing adequate instruc- and general and did not properly prepare the stu- tion in fundamental military skills and in the areas dent in the officership disciplines unique to the of career development and leadership responsibil- rated officer. The latter school argued the following ities, two areas which along with courses in com- factors that were revealed through the curriculum municative skills had for years represented the comparisons: bulk of officer training given to pilot trainees in • There is no common environment or time AIR FORCE REVIEW 75

element among the precommission schools, afrotc rather flexible program that could be adjusted to is conducted in hundreds of civilian educational fill the specific needs of a particular class or, institutions over a four-year period, ots is con- ideally, the needs of an individual. Accordingly a ducted in a military school environment over a “postgraduate” philosophy was adopted. Under three-month period. The Air Force Academy pro- this philosophy atc would establish the broad gram is conducted in a military school environ- guidelines for officer training in upt but the spe- ment over a four-year period. Thus it is extremely cifics of instructional material and the depth of difficult to achieve standardization in curriculum training pursued within these guidelines would and in the depth and quality of officer training vary according to the background and needs of conducted in the precommission schools. each individual class. • The curriculum of the precommission With this philosophy in mind, six general schools is tailored to cover the broad, overall objectives for officer training were developed, each aspects of officership. There is little emphasis on of which will receive detailed treatment: the unique aspects of the duties, responsibilities, (1) The student must be continuously exposed and career development of the rated officer. to exemplary standards of officership and must be • The majority of students entering pilot motivated to emulate this example. training (approximately 90 per cent) have had no (2) The student must gain an appreciation and previous commissioned service. knowledge of the concepts, tradition, and heritage handed down to him by his predecessors in the • Transition from civilian or noncommis- cockpit. sioned military status to commissioned officer status (3) The student must gain a knowledge of the creates problems of adjustment. Actual military specific responsibilities and additional duties iden- duty often differs from the expectations gained tified with his rated specialty. through theoretical exposure and preparation in a (4) The student must be given opportunities quasi-military environment. to develop leadership qualities and put previous • A large proportion of students entering training and education to use through practical pilot training have not committed themselves to application. career status. In some cases students lack a deep (5) The student must gain an introductory motivation to fly and have entered the program knowledge of the Air Force role in counterinsur- to “look it over" and see what it has to offer. gency operations. (This is a usaf requirement.) In light of these factors Air Training Com- (6) The student must attain and maintain a mand decided that continued officer training was high level of physical fitness. needed to prepare the student to assume the duties In developing the revised curriculum, special of a rated commissioned officer upon completion attention was given to the problem of duplication of the upt program. These factors also served as of training. Since the precommission schools offer the basis for developing the revised curriculum instruction in extremely broad subject matter areas, for officer training in the upt program. a small amount of duplication had to be accepted The lack of a common officer training back- in order to provide continuity and logical transi- ground among student pilots posed significant tion within the upt program. In the areas where problems in the development of the revised cur- duplication does appear, every effort was made to riculum. For example, classes entering training in point the instruction more specifically toward the July and August are composed primarily of Air junior rated officer. A paramount consideration in Force Academy graduates. Other classes are a curriculum development was to move from the mixture of ots and afrotc products. Interspersed general aspects of officership covered in precom- among all classes in training are a few officers with missioned training to more specific and narrow previous service, primarily navigators vying for areas directly related to junior rated officers. a dual rating, and a few Air National Guard offi- cers, many of whom have had little or no pre- commission training. leadership by example

This problem area dictated the need for a The young officers in the upt program are 76 AIR UNIVERSITY REVIEW normally very impressionable. Therefore one of the sistently exposed to exemplary officer-airmanship. most critical considerations in developing the cur- Each oti supervises a class of approximately 45 riculum was to provide effective leadership by ex- students. ample. As will be pointed out later, this leadership The oti’s function is extremely broad and factor is an essential element in making a “post- diversified. He flies instructional flights with stu- graduate" philosophy workable in a junior officer dents on a regular basis, provides classroom in- training program. struction in officer training subjects, monitors and The system of leadership by example con- often participates in physical training with his

The rigorous physical training portion of the undergraduate pilot training pro- gram is designed to get students into top physical condition and keep them there. ir>. v u In the physical conditioning phase stu- dents often wear “Elmers Handicaps.” l - / f KJ. * IN w

> i * - A 1, - ."V

bodied in the upt officer training program revolves assigned class, conducts inspections, counsels stu- around an experienced, knowledgeable, profes- dents as required, prepares Officer Performance sional instructor pilot assigned as the immediate and Training Reports on all assigned students, and supervisor of each class in training. This profes- is responsible for overall coordination of all train- sional is known as the officer training instructor ing activities for his assigned class. ( oti). While the idea of assigning a permanent- While there are few firmly established pre- party supervisor to each class is not new or unique requisites for qualification to serve as an o n , most to the upt program, it has been refined and devel- bases generally observe the following criteria: oped to a point where the supervisor /student rela- captain tionship is much closer and more comprehensive Senior Pilot rating than any observed in other programs. Through the one year minimum experience as upt in- oti system the students are immediately and con- structor pilot AIR FORCE REVIEW 77

outstanding military bearing graduate philosophy prevails in all formal class- graduate of Squadron Officer School room hours and in the many additional classroom college graduate hours devoted to special subject matters needed by communicative ability a particular class. Through his close association evidenced leadership ability with the class, the oti is able to analyze class needs voluntary acceptance of long, hard duty effectively and tailor needed instruction accord- hours ingly.® completion of Academic Instructor Course In summary, the oti function is to provide or equivalent (mandatory). instruction, supervision, and leadership by example for an assigned class of approximately 45 students. To be effective, the oti must be an experi- (Class size is programed to increase to 62 students enced. highly qualified instructor pilot and must in fy 66, which will make the oti job even more fly with each student in his class as often as pos- demanding.) By the time his class graduates, the sible. This is an essential factor in gaining the re- spect and confidence of the student and in pro- oti has been closely associated w'ith it for 55 weeks, during which he has observ ed, monitored, viding a means to gain insight into the student’s total personality, particularly his ability to react and influenced his students in all areas of training. in a pressure situation. It is important to note here that the students To ensure a complete knowledge of each stu- are not the only beneficiaries of the oti system. dent, the on remains with his class for the entire By maintaining a spirited, well-disciplined student 35-week program, which is broken down into body, the oti directly assists flight-line personnel three phases: Preflight. Primary (T-37), and Basic in accomplishing their mission more effectively. (T-33 or T-38). While flight-line instructors spe- Furthermore the challenging scope of his job is cialize in only one phase of training and in one tremendously broadening for the oti personally, type of aircraft, the on is with his class through all and the Air Force realizes the corollary benefit of the phases. He greets the class when it arrives, on-the-job training of officers destined for the more coordinates Preflight training, instructs in the Pri- demanding leadership positions. mary phase of training, continues to instruct his class in the Basic phase, and coordinates the grad- Heritage uation activities for his class. He is qualified in To instil] an appreciation of the concepts, both mission aircraft but maintains currency only tradition, and heritage handed down to the student in the one in which his class is receiving training. by his predecessors in the cockpit, a 14-hour course The oti normally does not cam' assigned stu- titled “Heritage” was developed. In this course dents on the flight line. Rather, he circulates students concentrate on the past and present lead- throughout the entire class, flying with each stu- ers associated with air power, their contributions dent as often as possible. Based on an average to air power, and the concepts that they helped to monthly flying time of 30 hours, composed of formulate. Since all precommission schools offer sorties of 1 hour IS minutes, the oti will fly 286 some instruction in Air Force history, no attempt sorties with his students during the course of train- is made to cover a detailed chronological develop- ing. This indicates that he will fly approximately ment of air power in terms of specific dates, events, seven times with each student in his class. At the and developments. request of flight-line supervisors, the oti will often The course is constructed around four broad fly consecutive missions with marginal students. periods: development of early air power, air power This is done to provide continuity of training and between the w'ars, air pow'er in World War II, and to capitalize on the on's flight-instruction experi- the development of modern air pow'er. Each period ence and personal knowledge of the student. °Thc OTI wife plays an important part in the program. Shortly after a class arrives, she invites the students' wives for A major portion of the o n ’s job revolves coffee, helps them get acquainted, answers their many questions around classroom instruction. The revised upt syl- concerning protocol, encourages participation in (he Officers Wives Club activities, and stresses the many ways they can help labus reflects three formal officer training courses: their husbands get through the demanding program. The OTI wife remains the primary adviser and counselor for the class Heritage. Duty, and Counterinsurgency. The post- wives throughout the program. 78 AIR UNIVERSITY REVIEW receives a similar treatment. First, the instructor ing man” concept—that the rated officer is a fight- conducts a one-hour lecture and discussion on the ing man sworn to protect the ideals of democratic major events and personalities associated with that existence, that he must stand ready to fight what- time period. The period involved is too broad to ever enemy the President may direct, and that he be covered in its entirety, and it is here that the must master certain responsibilities and disciplines postgraduate philosophy comes into play. The in- unique to his semi-isolated combat environment. structor guides the hour, feeling out the class for The first four hours of the Duty course are specific areas of interest. When the class responds devoted to a treatment of the freedoms that the to an event or series of events, the instructor builds rated officer has sworn to defend, the nature of the remainder of the hour around this area of ex- past and present threats to our freedoms, profes- pressed interest. He then assigns specific person- sionalism in the rated officer, the implications of a alities and concepts to the students for research commission as related to a rated officer, and a re- in preparation for seminar discussion. After two view of the Code of Conduct. or three days of preparation, a two-hour seminar Next the students are given an opportunity discussion of personalities and concepts is con- to discuss various instances of proper and improper ducted. In all seminar activities, classes are broken behavior of American fighting men, actual or hypo- down into small groups with a student leader. The thetical, in peacetime and in combat. This is ac- instructor monitors the seminar, taking notes and complished through a two-hour seminar in which observing the participation and communicative students discuss problem situations. The students skill of the students. At the termination of the have ample time to study the problems and to seminar, the instructor critiques the discussion and form definite opinions prior to the seminar. clarifies any unresolved points or vague issues. In keeping with the postgraduate philosophy, Through the study and discussion of past and typical situations are provided in the Instructor present Air Force leaders, students are led to an Guide, but instructors are encouraged to use situ- awareness that there is such a trait as patriotism ations or problems of their own selection so that and that valor and fidelity still have a place in this they can capitalize on the class’s interests and use age of sophisticated weaponry. One method of personal experience to lead the discussion to im- evoking student response to patriotism and valor portant points. All problems are locally reproduced is to relate typical Medal of Honor citations. No as student handouts. attempt is made to cover up or excuse the fact that The next important part of the Duty course mistakes in concept and judgment are also part deals with the specific duties of junior rated offi- of the Air Force heritage. By discussing these mis- cers in each of the commands. This unit of instruc- takes, the student benefits from the lessons of tion is designed to assist the student in selecting history. his assignment following graduation from upt. To encourage investigation and exploration, Typical data covered include checkout require- the Heritage course is nongraded. Thus instructor ments, typical mission profiles, typical additional and student have complete latitude for explora- duties assigned to junior officers, and the amount tion. No specific dates, times, and places need be of flying time the young pilot can expect to re- memorized just to be regurgitated for an examina- ceive in each command. One hour in this area is tion and then forgotten. This has proved to be a devoted to a panel discussion, using as panelists significant inspiration in developing interest in the qualified pilots who have recently returned from Heritage course. the various commands. The panel members are thoroughly briefed to be as objective as possible in their responses and not try either to sell or Duty downgrade a particular command. Students are To provide the student pilot with a basic encouraged to ask questions during this peiiod. knowledge of the specific responsibilities and addi- Although scheduled as a one-hour period, the tional duties associated with his rated specialty, a panel presentation invariably goes overtime and is 20-hour course titled “Duty” was developed. The usually continued at the Officers Club after duty emphasis throughout this course is on the “fight- hours. Also incorporated into the Duty course is a three-hour block of instruction dealing with rated officer career problems. Objectives of this block are to ensure that the student is aware of the impor- tance of the Effectiveness Report, the importance of education, the fact that each officer has an indi- vidual responsibility in career planning, and the major personnel policies governing his career— afm 36-1, afm 35-1, and afm 35-11.

Leadership Application The culmination of the Duty course is a series of practical leadership problems in which the stu- dent is exposed to both leader-assigned and leader- less group problem situations. Throughout the execution of the problems, students are closely observed for aggressiveness, communicative abil- ity', and cooperativeness. Frequently a student in the problem-solving group will be briefed to be an agitator or dissenter, thus complicating the group’s task in sohing the problem. Each problem-solving group is critiqued by a group of senior students who have previously accomplished the problems and have been detailed as observers. Thus a valuable by-product is realized in the form of student practice in analysis and critique. Another invaluable benefit inherent in the practical problem is that it provides an opportunity for the on to observe and evaluate his students in practical leadership situations. His observations are valuable later in counseling and in preparing Officer Performance Reports and Training Reports on his assigned students. The number of practical leadership problems executed depends on the size of the class and the consequent number of problem-solving groups. The maximum recommended group size is ten stu- dents, and each student normally participates in at least four different problems. The presentation of the practical problem unit is necessarily flexible because of varying facil-

Valuable experience and insight into student lead- ership ability arc gained by practical leadership problems incorporated in the Duty course. Senior students observe and take notes for the rritiipie as pilot trainees solve a "mined road" problem. 80 AIR UNIVERSITY REVIEW

ities at the different upt bases. Again, problems The student studies the guerrilla warrior, his back- outlined in the Instructor Guide may be used “as ground, motivation, training, and political ideol- is” or modified to suit local conditions. Problems ogy* require frequent revision to preclude compromise In this course students review some of the and common knowledge of the proper solution. tactics and operations employed in past counter- Perhaps the most important lesson learned by insurgency operations by comparing the insur- the student in this unit of instruction is the realiza- gency movements in Malaya and Indochina. This tion that problem recognition and a central source discussion leads into a study of general techniques of direction are of utmost importance. Invariably, and requirements of counterinsurgency operations. during the first problem or two, problem-solving The student also learns how U.S. Special Forces .groups waste a great deal of time haphazardly employ guerrilla tactics in combating the enemy. trying to solve the problem in a piecemeal fashion. The course includes a discussion on the Air After the first or second problem, however, the Force role in counterinsurgency operations. Typi- students realize that they must analyze the situa- cal items covered are the employment and special tion and determine where the real problem lies characteristics of coin air power, the aircraft types before taking any action. being used in coin operations, and those pro- The Duty course and Leadership Application gramed for future use. •constitute an exciting, stimulating experience for In a final two-hour seminar the student has both the^tudent and the instructor. It is an ideal an opportunity to review the course and express "vehicle for its designed purpose: to acquaint the his views on counterinsurgency. student with the specific requirements of his future Since the coin course is limited to 12 hours, assignment, to impress on him the individual re- no attempt can be made to provide a thorough sponsibilities in career planning, to remind him of understanding of the many complex factors and the ideals he has sworn to protect, and to give him actions that the subject embodies. Rather, the an opportunity to display and further develop his primary objective is to stimulate interest and en- leadership abilities. As in Heritage, no formal courage further study in this area. Student critiques grade, as such, is given. and seminar performance indicate that this objec- tive is being realized and that the young officers in pilot training find this to be a fascinating subject. Counterinsurgency

A requirement to indoctrinate pilot trainees physical training in the basic elements of counterinsurgency ( coin) operations was established by Headquarters usaf. The importance of maintaining a high state of This requirement fits well into the officer training physical fitness while undergoing a demanding curriculum in that it provides an opportunity to flying training program cannot be overemphasized. discuss warfare in terms of present-day problems. During the upt program students receive a mini- The course starts with a familiarization of the mum of 143 hours of scheduled physical training. history of insurgency and guerrilla warfare. The This includes 5BX indoctrination; conditioning student learns that guerrilla tactics are not new, exercises with and without wrist, ankle, and vest that they date far back in recorded history, and weights (called “Elmer’s Handicaps”); supervised that present-day emphasis on guerrilla tactics competitive sports; combative training, including stems from the fact that they are ideally suited to judo, boxing, and wrestling; swimming instruction insurgency movements. The student learns that and water survival techniques; and an obstacle insurgency is an increasingly valuable tool to the course consisting of a rope climb, parallel bars, in- Communist powers in their quest for world domi- cline wall, 30-foot tire lane, wall with cargo net, nation. balancing logs, horizontal bars, series of four-foot- Other objectives of the course revolve around high vaults, hand-over-hand, the dirty name, and the causes that underlie and inspire insurgency a jump-and-land. movements and an understanding of the differ- Three physical fitness tests are administered ences between conventional and guerrilla warfare. during the course. Designed to provide an easily AIR FORCE REVIEW 81

administered method of sampling a student’s over- attended by the rest of the student body. It is not all physical fitness, the tests consist of sit-ups, pull- uncommon for a keg of beer to be wagered on the ups, and a 300-yard shuttle run. The first test is outcome of a highly spirited play-off. administered during the 12th and 13th hours of In summary, the physical training portion of the physical training program, to let both student the officer training program in upt is aggressive and instructor know where the student stands at and fast-moving, designed to get the trainees in the beginning of the program. The second test is shape and keep them there. Corollary benefits are administered during the 60th and 61st hours of realized from the competitive spirit that the pro- training, immediately following the supervised gram generates, from the evaluation data provided team sports period. This is a graded test, the by formal testing, and from the oti’s being able to grade based on the number of pull-ups and sit-ups observe closely the traits of sportsmanship, co- and the time required to complete the shuttle run. operation, and competitiveness among his assigned The information gathered from the test is con- students. verted to a scoring scale and recorded on atc Form 99. Physical Fitness Record. The third test augmenting previous experience is identical to the first and second tests and is ad- ministered during the 90th and 91st hours of physi- As mentioned earlier, students entering upt cal training. have been exposed to various degrees of previous The physical training tests are conducted out- officer training. The postgraduate philosophy in doors. Consequently, temperature and weather the upt officer training program offers many op- variances on testing dates influence the results. portunities to extend this previous experience in Nevertheless the test results do reveal some inter- several ways: esting albeit not rigorously defensible data. —by assigning existing local problems to For illustrative purposes, average physical students for study and recommended solutions. fitness test results of classes 64A and 64B, primarily Students are thus actively encouraged to formulate Academy input, were compared with the average well-thought-out solutions to problems. Many results of classes 64C through 64H, a mixture of worthwhile ideas have been realized through this ots, afrotc, and previous-service personnel in- procedure. puts. These classes were undergoing training at —by allowing students to plan and coordi- Webb afb in fy 64. Test results indicate that all classes progress significantly in their ability' to per- nate parade, retreat, dining in, and graduation form the three test items, an apparent indication ceremonies. Although the cm acts as overall co- that the students’ general level of physical fitness ordinator for these affairs, he delegates the majority also improves during the course of training. Inter- of the tasks to the class and evaluates individual esting to note is the fact that the Academy classes student initiative and ingenuity in their accom- scored higher on initial testing in the pull-up and plishment. shuttle run items than the ots and afrotc classes. —by assigning students to prepare and pre- This may indicate that the Academy physical train- sent briefings on subjects of interest. Again this ing program stresses upper extremity and running area, revolving around the oti’s ability to deter- exercises more than the afrotc and ots programs. mine the needs of a particular class, is in keeping The Academy classes achieved a better final per- with the postgraduate philosophy. This is func- formance in all three test items. tional activity in every respect: it provides an op- Besides providing a method of maintaining a portunity to practice and evaluate communicative high degree of physical fitness, the physical train- skills, brings items of interest to the attention of ing program provides an ideal method of instilling students and instructors, and uses to great advan- a competitive spirit among the students. Through tage what might otherwise be nonproductive time, the competitive sports program, classes compete e.g., weather days. directly for honors. Class standings and won-and- —by establishing and observing arf organi- lost records are posted daily. After-duty play-offs zational structure resembling a squadron within in case of ties are fairly common and are well each class. Normally the ranking officers are ap- 82 AIR UNIVERSITY REVIEW pointed to the command positions in the class; dered twice on each student—after the 24th week however, Webb afb is enjoying great success with and after the 48th week of training. The form a system in which internal class command positions used, atc Form 332, closely resembles af Form are periodically rotated among the more junior 77, Officer Effectiveness Report. Through its use officers, thus giving them an opportunity to gain the oti converts descriptive statements of behav- leadership experience. ioral traits to a standard score grade as explained in In these ways the student body’s exercise of atc Manual 50-4, Improvement of Grading Prac- initiative and imagination enables both the upt tices in ATC Schools. Data from the Officer Per- program and the students to benefit from previous formance Rating are also used by the oti to pre- leadership experience. pare student final Training Report, af Form 475. To prevent inflated ratings, the oti is required to maintain an average raw score of 21 for his evaluation class. This is the average, straight-down-the-middle A formidable problem in developing the re- rating, and when converted it equals a standard vised officer training program arose in the area of score of approximately 85. A standard score of 77 evaluation. While paper-and-pencil tests provide or below is considered to be failing. This equates a practical method of gathering evidence about to a raw score of approximately 17 or below on certain types of student behavior, the behavioral current conversion scales. A student receiving a changes sought through officer training were con- failing opr grade is given a reasonable amount of sidered too intangible to be appraised validly time, six to eight weeks, to demonstrate improve- through that medium. The overall objectives of the ment. If he fails to show satisfactory improvement, officer training program represent personal, social, he is recommended for military elimination from and idealistic adjustment, and therefore they are the program and for consequent action under afr realistically appraisable only through actual obser- 36-2 or afr 36-3. vation of student behavior in actual performance The standard scores rendered on the two situations. Officer Performance Ratings constitute 80 per cent The solution to the problem of evaluation of the student’s final officer training grade. The stemmed from the belief that the students should remaining 20 per cent is obtained from perform- be evaluated in the same manner that would even- ance results in physical training (10 per cent) and tually be used to evaluate them in the field- formal evaluation in the coin course ( 10 per cent). through performance or effectiveness ratings. How- The final officer training grade constitutes one ever, the overall standards of an evaluation method third of the student’s final grade for the upt —objectivity, reliability, and validity—were matters course. The remaining two thirds of the overall of serious concern and not easily resolved. grade are derived equally from performance in fly- Two primary criteria for evaluating officer ing training and academic training. training were established: The subjective evaluation system designed ( 1) The evaluation system must appraise actual for the revised officer training program is a rather student behavior, not the mere ability to recall radical departure from standard grading proce- facts and figures. dures and practices in most other training programs. (2) The evaluation system must incorporate It was designed to be compatible with the post- frequent appraisals in order to identify changes graduate philosophy of the program and to impress that may be occurring. on the student that his “grade, so to speak, when The oti, through his close daily association he gets out into the field will be his effectiveness with the students, was the logical agent to appraise report. The method of grading and the weight actual student behavior. To increase the objectivity given officer training grades underscore the im- of his rating, the oti solicits and compiles addi- portance of effective officership and emphasize the tional observations from flight and academic in- underlying notion that the development of officer- structors. ship traits is paramount in the final analysis of Officer Performance Ratings ( opr) are ren- Air Force needs. AIR FORCE REVIEW 83

T here is a definite need for officer training in the Honor, Country—are the primary ingredients of the undergraduate pilot training program despite the new revised officer training program in upt. fact that student input consists of commissioned Through the postgraduate approach to officer officers. The revised officer training program devel- training, Air Training Command is realizing the oped for the upt program was designed to em- exciting, stimulating experience of instilling a love phasize the need for officer quality, to provide an of service and duty and a quest for professionalism appreciation of Air Force heritage and tradition, among the fledgling rated officer force. As long as to qualify students for effective performance in the program achieves this purpose, the time and their postgraduation assignments, to increase stu- effort expended in continuing officer training will dent awareness of the Air Force role in counterin- certainly be worthwhile. surgency operations, and to maintain a high degree 3360th Pilot Training Wing of physical fitness among students undergoing training. In the pursuit of these objectives, emphasis is placed on the spirit of patriotism, valor, and Acknowledgment fidelity—traits too often dismissed as being old- The author, as Chief, Officer Training, Webb AFB, had the assistance of the Officer Training Instructors at each under- fashioned in this day of modem, sophisticated graduate pilot training base and at Janies Connolly AFB in weaponry. Leadership by example and continuous constructing the new UPT course. Captains James W. Marlin, Jr., and Robert H. Laurine of Webb AFB made significant con- reference to the three meaningful concepts—Duty, tribution to the preparation of this article. AIR FORCE OFFICER EDUCATION

C olonel John P. L isa c k

NUMBER of important aspects of the Air cation affects the quality of both consciousness and A .Force officer corps are slowly but surely behavior. changing. One of these is officer education. Not Broadly speaking, officer education may be only is the career of every officer affected directly considered in two categories—specialized educa- by his educational attainments, but the very nature tion and professional military education. I shall and effectiveness of the Air Force is involved. describe each briefly and then discuss both in some Although I shall treat essentially the special- detail. ized education and the professional military edu- Specialized education in the Air Force is es- cation needs of the Air Force officer, at least four sentially civilian college-type study that relates to distinguishable elements should be present in the each Air Force specialty. These studies normally education of any professional: take place in an accredited college or university. • In order to fulfill his civic duties as a The Air Force Educational Requirements Board citizen and live an enjoyable, informed, and effec- has identified and described the educational levels tive personal life, he must have the necessary and appropriate fields of specialized study that are general education. needed for full qualification for each officer af specialty. Further, the Officer Classification Man- • His education must include the broad ual, afm 36-1, has a paragraph describing the theoretical base required for his professional prac- broad minimum educational requirements for each tice. specialty. In other words, specialized educational • He must have mastery over the special- requirements have been fairly well described and ized knowledge and technical skills which char- recorded. acterize his profession. (In this regard military Specialized education is largely given to af activities are unique; they are not solely techno- officers through the formal resident programs of logical or mechanical nor are they commercial, the Air Force Institute of Technology at Wright- societal, or the like. Hence the military profession Patterson afb, Ohio, or through civilian universi- is a distinct one.) ties under afit contractual arrangements. In addi- 0 He must keep mentally active continu- tion thousands of Air Force officers are engaged ously in both the general and professional cate- in specialized educational programs through Boot- gories. He may do this on a formal or informal strap and “on their own.” basis, either individually or in a group educational A project now nearing completion defines, program. categorizes, and codes all specialized educational It can be readily seen that the Air Force policy terms. Very soon each officer’s education will be of procuring officers initially with the bachelor accurately reported, and by means of electronic degree and providing subsequent specialized and data-processing techniques his specific educational professional military educational programs is in attainments, by level and field of study, can be consonance with and supports these professional quickly ascertained. This project will make spe- qualifications. The Air Force recognizes that edu- cialized education more important and meaningful AIR FORCE REVIEW 85

in personnel and manpower management actions. a specialist, particularly in some of the scientific On the other hand, professional military edu- fields, leaves little or no time for military educa- cation ( pm e ) is made up essentially of the educa- tion. Further, they may ask: If the officer is em- tion required for the development of a professional ployed as a specialist, why does he need profes- military officer regardless of his specialty. Pro- sional military education? grams that provide this education pertain largely A basic conflict does not exist because every to the military arts and sciences. There is neces- officer is necessarily both a specialist and a pro- sarily some overlap with what are normally con- fessional military man. There is no officer who sidered to be specialized educational subjects be- does not have an Air Force specialty that requires cause the profession of arms includes consideration specialized skills (and most have a number of of many of the well-recognized collegiate-type them), and there is no man or woman w'ho is a fields of study. Military science builds on these lieutenant or captain or field-grader who is not specialized studies. For example, selected studies an officer too. Frequently where specialist skills in the social sciences, engineering and physical alone are needed, civilians supply them or con- sciences, administration and management, to name tractual arrangements are made. In any event a few, are elements of military science. However, sufficient numbers of active duty and reserve mil- the broad context in which all pm e studies are itary personnel must be available to progress to made relates primarily to national security from the higher military officer positions, and sufficient a military standpoint, and the specific educational numbers must be trained and ready to meet war objectives pertain essentially to the continuing de- plan requirements. While it is true that some offi- velopment of a professional military officer, pm e cers are engaged more in specialized work than is therefore necessarily conducted in schools and in general military duties, both elements are al- colleges controlled and administered by the mili- ways found to some degree in every position or tary. For usaf officers formal pm e studies begin job filled by an officer. It is also important to with military subjects in the officer precommission- realize that these relationships may vary as an ing programs, extend through the Squadron Officer officer progresses in his career or serves under war- School for lieutenants and captains, the Air Com- time conditions. Thus it can be seen that a “con- mand and Staff College for captains and majors, flict exists only if one is determined to cast all and the Air War College for lieutenant colonels. officers in a rigid mold and retain the casting con- The military staff and war colleges of our sister figuration indefinitely. services and the joint military colleges (such as The changing nature of an officer’s responsi- the Armed Forces Staff College, the Industrial bilities is indicated in Air Force Regulation 36-23: College of the Armed Forces, and the National Managerial responsibility is inherent War College) broaden and top off the pm e system within each officer position. Normally, it in- of the U.S. armed services.0 Allied officers share creases in proportion to the officer’s progression in many of these U.S. programs, and reciprocal to higher and more responsible positions. In arrangements have been worked out for represent- senior field grade positions, managerial ability atives of our semces to attend pm e schools of our usually outweighs the requirements for specific allies. technical job know ledge, except in some of the Many writers have claimed that a basic con- most highly complex specialties such as chem- flict exists between the two officer educational ist, metallurgist, weather officer, procurement categories, the contention often being based largely officer, and intelligence officer. on the argument that large numbers of officers Officers will be relatively narrowly spe- cialized in the early years of their careers to are required as “specialists' and not as “military develop the technical competence required and generalists. They claim that keeping abreast as to provide a sound foundation for further career development. This is essential for those who °Tbe Air Training Command is responsible for conducting will further specialize, and serves as the base pilot and navigator training programs and technical training courses to meet specific officer job requirements. Although these on which to provide broadening for those pro- include study of subjects which are in the area of PME, they are gressing towards the status of qualified senior separate from the two major categories of specialized and pro- fessional military education and are not considered here. officers. Broadening utilization must be com- 86 AIR UNIVERSITY REVIEW

mensurate with the officer’s increase in experi- accomplishment be identified and achieved and ence, grade, qualifications, and capability. the officer be utilized accordingly. The latter re- It is a fact that even in the scientific specialties quirement dictates that positions calling for gradu- many officers who do research at squadron officer ate-level education be identified throughout the level are called upon to direct and manage re- Air Force. search and development activities at the field- In order to view the specialized educational grade officer level. As the management level in- characteristics of Air Force officers in proper per- creases, more military considerations are neces- spective, one must look back to World War II when sarily included. enormous numbers of young men who had little or no formal college education were trained as pilots and navigators through the Aviation Cadet specialized education Training Program. This program has continued An officer’s specialized educational attain- since the war with little change up to recent times, ments are depicted in large part by the earning and a large segment of the present officer corps of an academic degree at the bachelor’s, master’s, was procured through it. The Aviation Cadet Pro- or doctor’s level. The general areas of study and gram officer input, compared with all other officer major academic fields, from accounting to zool- procurement programs of the past, is reflected in ogy, are usually identified with an academic de- the present composition of the officer population gree and are fairly well known. The knowledge as shown in Figure l.° and understanding acquired through studies in certain of these academic fields are directly appli- cable to specific Air Force specialties. They help A N im m e d ia t e generalization about the officer to become proficient in the duties and officer education can be made from this chart: responsibilities unique to the specialty, such as that only about one third of the present officer accounting for a comptroller or meteorology for a corps has come from precommissioning programs weather officer. It is always important—but even more important under austere conditions—that the “Data for this and following charts were obtained from USAF Officer Master File as of 31 May 1963. All Air Force type and level of education essential to mission officers are included.

No. Officers Per cent

Cadet Appt fr Mil Appt fr Civil Acad Mil Figure 1. Officer popula- te serv tion by source of comvussion AIR FORCE REVIEW 87

which require a college degree to qualify for the which culminates in the award of a college degree, program or that resulted in the award of a degree has replaced Officer Candidate School, and only upon completion. These sources include the graduates of civilian colleges or the academies are afrotc, the academies, and the Officer Training going into pilot training. Further, it is noteworthy School. As a result, until a few years ago, less than that the number of cadets in the Air Force Acad- half of the line officers in the Air Force had emy is being increased. achieved a college degree. Educational Level by Years of Service and The minimum educational standards applied Grade. The results of the older officer procurement in World War II remained in effect for officer pro- policies, most of which did not emphasize educa- curement programs for nearly 15 years after the tion, and the newer ones which do, are vividly de- end of th'e war. Then came a change. picted in Figure 2. It is clear that a larger per- An important personnel goal, namely, that centage of new officers have bachelor’s degrees. each officer should have at least a bachelor’s de- Also, the more years’ service officers have, the gree. has been revived through the increasing fewer of them have college degrees. recognition of the need for adequate educational Because of the preceding disclosure and the preparation in all areas of aerospace endeavor. fact that an even higher percentage of newly pro- This recognition has been abetted by the specific cured officers have degrees, at first blush it appears findings of the Air Force Educational Require- that an anomaly exists: that senior-grade officers ments Board, stimulated by the pressures of in- have lower educational levels than their juniors. creased individual performance requirements and However, such is not the case. The fact of the advancing technical requirements, and further matter is that despite the lower average education jolted by the competition of foreign powers in level of each sequentially older group, the better- space and modem weapon systems development. educated officers as a ride have been the ones who Implementing actions in support of this new were promoted over the years. The generals as a policy are having a direct and measurable effect on group have a higher educational level than the the composition of the officer corps and its special- colonels, and the lieutenant colonels and majors ized educational levels. Line officers (except for a are each sequentially lower. Since new officer small number in navigator training) are not now- procurement policies emphasize education, a high being procured through any program unless they percentage of all lieutenants have degrees, and the already have a degree or are awarded one upon higher proportion is beginning to show at the cap- completion of the precommissioning program. The tain grade. The officer education levels are shown Airman Education and Commissioning Program, by grade in Figure 3.

Figure 2. Officer educational Years of Service 20+ 15-19 10-14 5-9 0-4 levels shown by years of total n □ active Federal military service Doctor Master Bachelor 100%

90

m m □ Figure 3. Officer educa- Doctor Master Bachelor tional levels by grade

It is clear that a selection system has been at problems of the aerospace age which they will work resulting in promotions to field grades for face. This specialized educational base—in the those with college degrees despite the progres- proper fields of study—will be the foundation upon sively lower average educational levels of the which to build those skills needed for the increas- older groups of which they are a part. It might ingly complex duties and responsibilities. be argued that the reason generals and colonels The Air Force Educational Requirements have a higher formal educational level is that they Board ( aferb) has clearly established that to have been in the service longer and have had more meet future needs higher levels of specialized time and opportunity to advance their education. education will be required. This should surprise If this contention were true the down line pattern no one, for a like trend is evident in most civilian shown in Figure 2 would then be reversed, or at professions and vocations. Because things are mov- least it would resemble the saddle contour in ing ever faster, the job of keeping abreast of and Figure 3. Another interesting fact is that there is contributing to the advancements in his chosen a positive correlation between the higher-level de- field will become a greater challenge to each officer. grees and the higher grades. No doubt many readers have determined just The Air Force will suffer heavy losses over where they fit into each of the charts. I am con- the next few years from the retirement of officers vinced that no matter where it might be, each offi- with World War II experience. It appears from cer can rest assured that if he keeps standing still, Figure 3 that officers in the middle grades, who the chances are that he will soon be left behind. will inherit the authority and responsibilities of This is borne out by a look at the past record and those now in the higher grades, must prepare now a look into the future, for which some computa- for the opportunities and duties soon to be thrust tions and predictions can be made. (See Figure 4.) upon them. Education is a key element of that Note the progress made in 1950 and 1951. This preparation. is due in large part to the recall of reserve officers The younger officers are moving in a circle to meet the Korean crisis, many of whom took of relatively well-educated peers. The data in these advantage of the C.I. Education Bill and went to charts show that their situation is quite different college after World War II. The emphasis on from that of their senior officers. The younger offi- education in the 1960’s is also beginning to show. cers have a better formal college educational base. The Sample Survey of Air Force Officers as of These officers will undoubtedly be better prepared March 1964 indicated that over 56 per cent of to meet many of the more advanced specialized all officers had achieved a bachelors or higher de- 100°o

Me 4. Percentage of Air Force ers with college degrees (actual projected) as of February 1964

gree. The future trend reflects the attrition from on a limited basis for a few officers. The pressing the force of officers procured through the sources need of maintaining a high state of operational of years past, the effects of new personnel policies readiness continues to provide closely related ex- and goals, and the unprecedented and laudable periences for others. efforts on the part of thousands of officers now In this discussion of education for military improving their educational levels. professionalism, the raison d’etre and nature of the organization are relevant. Therefore, the specific provisions of the law are pertinent as set forth in professional military education U.S. Code 10. Section S062, which reads, in part, An officer’s professional military educational that the armed forces must be capable of attainments are most frequently identified by not- (1) preserving the peace and security, and ing which pm e programs he has completed. This providing for the defense, of the United States; practice is probably due in part to the habit of (2) supporting the national policies; measuring specialized education by the degrees (3) implementing the national objectives; and awarded. In fact, some writers equate the Air (4) overcoming any nations responsible for ag- Command and Staff College and Air War College gressive acts that imperil the peace and securit> with the master’s degree level. Their position is of the United States. somewhat reinforced by the recommendation of Further, the oath of office states clearly that the American Council on Education to award some each officer “affirms to support and defend the graduate-level credits for successful completion of Constitution of the United States against all ene- these two top Air Force pm e college programs. mies, foreign and domestic . . I believe that some equivalence of military There is little question that the intent of the professional education has been acquired by offi- law and the oath of office is military oriented. One cers through the broad variety of assignments ear- can scarcely argue that some of the persons who lier in their careers and through direct military elec t to become members of the armed forces and combat experience. However, specialties have lie- take the oath are specialists (civilian type) and come narrower and assignment variety has less- not military men. ened. and there has been only little chance for It has been stated earlier that the level of combat experience since that great war in the 10’s. actual military-professional knowledge, the mili- Korea, the Cuban crisis beginning in October 1962. tary combat experience of officers, and the need anil now \ iet Nam have provided this experience for professional military education vary widely. AF Officers Still on Active Duty W ho Have Attended Professional Military Education Schools or Colleges

School or college Number Percentage Percentage of who of AF off. class who attended* who attended* are reg AF National W ar College 305 .23 99.67 Industrial College of the Armed Forces 352 .27 93.75 Air W ar College 1243 .95 97.67 Arm y or Navy W ar College 263 .20 89.35 Armed Forces Staff College 1296 .99 96.06 Air Command and Staff College 9942 7.58 82.73 Other command and staff colleges 1712 1.31 87.91 Squadron Officer School 32,545 24.82 57.66

Total number enrolled 47,658 36.35%* 66.77% Total number who have not attended 92,990** 70.93

Totals 140,648** 107.28%**

Actual officer strength 131,101

*Sinee the USAF Master File lists only the three highest lower levels because some officers may have attended schools attended, the numbers represent enrollments rather more than three schools (although this is not too common).* than individual officers; the numbers are incomplete at the **Multiple enrollments therefore account for the addi- tional 9547 officers (7.28%) over actual officer strength.

The required pme programs are therefore under- officers are members of the armed forces, and de- standably very difficult to define, pme involves terring war or conducting it is their business. the acquisition of knowledge, skills, and attitudes The reason for a continuous educational proc- that are requisite to military professionalism, and ess is to ensure that an officer’s professional military the knowledge and experience of the officers con- understandings are in line with his higher military cerned must be considered, pme is also needed to grade, which determines his authority and respon- ensure that a professional military officer realizes sibility. It was said that specialized education helps that he is engaged in war even as he strives to the officer to become proficient in the duties and deter it. He must be conditioned to expect war responsibilities of his specialty; pme helps him to at various scales and, whatever its intensity, be become a more proficient officer and increases his effective in his role. All officers are associated to knowledge in the military arts and sciences. The some degree with either present or future combat learning objectives of pme at lower level usually or combat supporting units, and each manages help the junior officer to acquire understandings resources or monitors activities related to the mis- concerning military subjects pertinent to lower sion; each should understand and be effective in military organization echelons and facilitate his his commanding, directing, planning, operating, or ability to apply this knowledge meaningfully. As supporting roles in peace and in war. Further, he advances in his career—gains experience and each will do his particular jobs more effectively maturity—he may be selected to attend more ad- when he understands how they contribute to the vanced pme studies. The depth of learning then achievement of the primary Air Force missions and increases, to improve his ability to synthesize and how these support national security objectives. We evaluate data concerning military and national cannot afford to have military officers on board security problems of broader scope. Not all officers who picture themselves as civilian-type specialists require the most advanced pme programs, and not who are spectators to the acts of deterring war all attend. In fact, as indicated in the accompany- or conducting a hot war once it has begun. All ing chart, less than two per cent of Air Force offi- AIR FORCE REVIEW 91

eers now on active duty have completed a war fessional side, af pme can and should help meet college-level program. The largest percentage of this deficiency. officers who attend the more advanced programs are regular Air Force officers. Only at the squadron officer level do the reservists attend in any large T he educ a t ion of each officer must be both realis- proportion. tic and idealistic—the former because it must meet All officers are professional military men. present and near future needs, the latter because Therefore, all need a military professional base, it must help to develop each officer to his fullest which must be updated continuously. Not all offi- potential. Officers who have argued for either spe- cers need the most advanced pme . nor do all have cialized or professional military educational pro- or need the characteristics to excel as "military grams must realize that neither one can be neg- tigers.” lected for himself or for those he leads. They are Completion of the first-level af pme (Squad- both necessary in the career development of a ron Officer School) is now required for all career professional military officer. The late philosopher officers, by correspondence if not by the resident Alfred North Whitehead said it well: program. Many officers believe that the broad In the conditions of modem life the rule is military experiences under combat conditions of absolute: the race which does not value trained the bulk of the present field-grade officers have not intelligence is doomed. Not all your heroism, yet made it necessary to post a mandatory require- not all your social charm, not all your wit, not ment for officers to complete the second-level af all your victories on land or at sea, can move pme. 1 predict, however, that the time is nearing back the finger of fate. Today we maintain our- when such will be the case. selves. Tomorrow science will have moved for- .Although junior officers are better prepared ' ward yet one more step, and there will be no in specialized education and become proficient in appeal from the judgment which will then be their specialty relatively early in their careers, they pronounced on the uneducated. are less well experienced on their military pro- Air Force Educational Requirements Board

A Description of the Professional Air Force Officer

T hf. professional Air Force officer is the aerospace expert of the Nation’s fighting forces. He understands the nature of war and is proficient in the art of waging it under any level of conflict. He is a leader of men in both peace and war, and he is accomplished in utilizing his knowledge and skills in organizing and managing resources. He combines military bearing and self-confidence with loyalty, integrity, self- discipline, versatility and adaptability. His conduct and ethics are based upon the ideal of service above self. He communicates effectively and works efficiently with people at all levels and from all walk' of life. He participates in specialized education, as well as specialized training, and he employs this continuing preparation, in conjunction with professional military education, in order to be able to assume greater responsibilities as lie progresses in his military career. The professional Air Force officer recognizes that he must continually expand his knowledge and understanding of the art of war. He recognizes his responsibilities to the Nation, and he thus seeks to maintain those high intellectual, ethical and physical standards requisite to a corps of professional officers which merits the trust and respect of the society from which it draws its authority and which it is duty-bound to defend. Extract from the AFERB Report on Professional Military Education, Volume 1, 10 July 1963, as modified by Hq USAF. LAST OF THE TEXAS TOWERS

Tow lines attached, tugs prepare to move away from Texas Tower No. 3 prior to the blasting of its legs. The explosive blast (belotv) demolishes the structures left leg. It was blasted a split second before the others to “launch” the struc- ture into the ocean at a slight angle. A helicopter lands crewmen on the 1600-ton structure as it is towed to Kearny, New Jersey. Its lower deck is welded shut and filled with urethane foam for flotation.

N 6 AUGUST 1964 the last of the Texas of making four gallons of fresh water per minute, O Towers. Tower Number 3. was plunged into and an 82,000-gallon tank ensured an ample the sea by a commercial salvage firm. The giant supply. radar stations, nicknamed for their likeness to off- The towers were divided into four main decks. shore oil-drilling rigs in the Gulf of Mexico, were The bottom one was mainly used for utilities; the erected on the continental shelf off the northeast second was living quarters; the third was an opera- coast of the United States between 1955 and 1957, tions area; and the top one contained the radar to plug gaps in the North American early-warning domes. Originally the towers were manned by system. One tower collapsed during a storm in 105 men, but automation eventually reduced the January 1961, with a loss of 28 lives. The other two requirement to 65. They usually served 30 days functioned 24 hours a day, seven days a week, aboard a tower and then 30 days on shore duty. until June 1963, when the Air Force declared that Helicopters and ships maintained contact with the long-range aircraft and other advances adequately towers. Ample electronic equipment assured the filled the original gaps. men of rapid communication with Air Defense The towers were in the shape of equilateral Command installations. triangles measuring 200 feet on each side. Sup- After the towers were declared surplus, an ported by three tubular legs 14 feet in diameter attempt was made to salvage Tower Number 2, sunk 45 feet into the ocean floor, the towers rose located 153 miles off of Cape Cod, Massachusetts, 87 feet above the water. The legs, each with 2,2 by blasting the legs, but it sank when it hit the feet of concrete between outer and inner steel ocean. Tower Number 3, positioned 30 miles tubes, were utilitarian: two were containers for southeast of Nantucket Island, Massachusetts, was fuel oil and one for seawater to be distilled for blasted from its moorings and salvaged by the Lip- I drinking purposes. The evaporators were capable sett Division of Luria Bros. & Co., Inc., both sub- 94 AIR UNIVERSITY REVIEW sidiaries of the Ogden Corporation, under super- times in volume to fill the space, adhered to the sur- vision of the U.S. Army Corps of Engineers. Lip- face, and set in minutes to a high-strength, moisture sett previously had dismantled the French liners and fire resistant, rigid mass. Normandie and Liherte, the battleships New Mex- When the foaming operation was completed, ico, Iowa, and Wyoming, and the aircraft carrier acetylene torches partially cut the legs 3 or 4 feet Enterprise, but there were some unique problems below the platform, explosives were set on each in this project. leg, and heavy charges were planted inside the Prior to the salvage operation AF Ground legs below the ocean floor. The explosive charge Electronics Engineering-Installation Agency crews on one leg was detonated a half-second before the dismantled the electronic gear and other equip- other two, to cause the platform to tilt as it dropped ment amounting to 800 tons. to the water. Oceangoing tugs had lines already To prevent the tower from sinking, the lower attached and towed it to Kearny, New Jersey. The deck was filled with a special lightweight urethane salvage firm may convert it into a floating dock foam, applied by the Poly Systems Division of equipped with a machine shop or may strip it of Dayco Corporation. This foam, formulated from the steel for scrap metal. In the latter event the ure- basic resin and isocyanate materials, increased forty thane foam will be sold for flotation material. TUNNER AND THE SA G A OF AIRLIFT

Colonel Raymond L. T owne

F GENERAL TUNNER were running the lion- usually penetrating quality of his displeasure. I taming act for Bamum and Bailey, the lion The jacket of his absorbing book. Over the would put his head in Tunner’s mouth—so watch Hump,° describes him as “the Man Who Moved . » out. Anything Anywhere, Anytime.” This description was confided to me by a Both descriptions are incomplete. freshly scarred colonel that day in 1946 when I In the years following the colonel’s warning, reported for duty in Tunner’s Memphis head- Tunner moved many of us physically all over the quarters: Continental Division, Air Transport Com- globe. He also moved us emotionally: from cold mand, rage, through sheer exhaustion, to the peaks of By this time Brigadier General Bill Tunner quiet pride in having hacked another “impossible” had already completed four years in the mass air- job. lift business. Behind him were the successful war- What was General William Henry Tunner time Ferrying Division and India-China "Hump” really like—and how did he do what he did? airlift operations. Ahead of him lay Berlin, Korea, How did he save Berlin from the Soviet Lebanon, Formosa, Big Slam, and a host of smaller strangle? but equally intense emergency airlift operations. How did he move four Chinese armies over But even by 1946 the Tunner legend had the Himalayas when Hannibal couldn’t move one crystallized around a man described as hard- winning army over the Alps? dri ving, coldly brilliant, and known for the un- How did he keep the U.S. Eighth Army going "William H. Tunner, Lt. Gen., USAF (R et), Over the Hump (New York: Duell, Sloan and Pearce, 1964, $6.95), 332 pp. plus index. 96 AIR UNIVERSITY REVIEW with rations, gasoline, and ammunition in Korea ers complained at being “Schuffertized” when their when there were no roads or rails and shipping mess halls were portrayed as serving roast camel, was out of reach? or as having gorillas cooking food “untouched by How did he inspire men who cursed his eye- human hand.” But it did them no good. Only the balls on the Hump operation—and flocked to his General could censor any of Jake’s masterpieces. command when the Berlin Airlift began? And then He has yet to censor the first one, except to keep did it all over again when the Korean Airlift began Jake’s “latrine humor” under control. to form. Since Jake’s works were printed in the official How did he do it? On balance, I believe he headquarters daily newspaper, every coal heaver, knew his mission better than any man in his com- nut twister, and airman knew that his problems mand, and he knew his men better than they knew were known to The Old Man even if he couldn’t themselves. For the professional officer and non- do anything about them immediately. com this is the most compelling lesson of Over the If leadership begins with knowing your men, Hump; and this is the lesson which must have in- then Over the Hump is filled with such lessons. It spired General Lucius Clay’s urging every military is filled too with Tunnel’s incisive “cause and man to read the book. effect” logic—the logic of a man who, even in the Item. Tiinner would not sit through a briefing heat of World War II. viewed his airlift mission that did not begin with a statement of the mission. not in static relation to organization charts or (“Don’t tell me what you’ve got—tell me what tables of equipment but as one of the dynamic you’re here for.”) forces soon to reshape our postwar world. Certainly by 1947 he saw the postwar world Item. Tunner would not sit in his office, or with a clarity of mind that, unfortunate when it any subordinate commander’s office, if he could happened, at times exceeded that of his superiors. be out flying with his men, or listening to their “Are we to become the Roman Empire of the gripes, or looking around to see what cartoons or Twentieth Century? Are we Americans, together slogans they had pinned up on their walls. (“Let’s with our allies, going to be forced to man the get out and see what the boys have got on their frontiers of the free world against the Communist minds today.”) Huns and Vandals’?” Item. Tunner was the most avid pin-up fan in He posed these questions to me one afternoon the Air Force. No, not for photographs of polka- at our Fort Totten headquarters of atc’s Atlantic dot bikinis. Tunner liked the one Captain Martin Division. By 1947 the hot war had been over less Luther tacked up in the Flight Operations Center than two years. But already chill winds were blow- one rainy night. This was after Luther had been ing off the steppes and Tunner was trying to sense told that he would have to work a double shift what this would mean to all Americans, and par- before he flew his daily load of coal into Berlin. ticularly to their military services. The pin-up was a plain piece of paper on which “If this is to be our future,” Tunner continued, Luther had neatly lettered: “Tons for Tunner: If “then we are going to need airlift forces of a size you don’t have a double hernia—you’re not pulling and capability we can’t conceive of today. How your share of the load!” else can one nation lead its allies in policing the When the General saw the embarrassed world except through a superior technology that Luther’s handiwork, he merely smiled and asked will give our military forces unquestioned flexi- him how the operation was going. Tunner knew bility and mobility?” his man. He knew that Captain Luther might Was he right? The answers in Over the Hump moan and groan. He also knew that Luther would —and your daily newspaper—are abundant. rather suffer the double hernia than let the mission It would be difficult for any military man to down. put this book down unfinished. Regardless of his Item. Sergeant Jake Sehuffert was ordered to assignment, airlift today touches the life of every cartoon the hardships, discomforts, and tribula- man who wears the Navy’s blue, the Army’s green, tions of the men who crewed and flew the Berlin or Shade 84. and Korean Airlifts. Naturally the base command- The F-104 pilot is aware that his ability to BOOKS AND IDEAS 97 continue to fly hinges on the airlift of his engine The trooper in the green beret sloshing replacements to Europe or Formosa or wherever through Viet Nam can get what he needs from the his slice of sky may be. zi when he needs it, through airlift. Or he can get Thule is no longer at the end of the world. stateside medical care within 12 hours if a Viet It is now less than six hours from Broadway by Cong bullet cuts him down. ma ts C-135. Another winner is the worldwide logistic sys- The Polaris expert tending his missile under tem of the Air Force. Indeed, the logistic systems the turbulent North Atlantic knows that he and of all the military services are vastly more effective, his fellow submariners will soon exchange places at far less cost, because Tunner’s principles of air- with a fresh crew flown in to a rendezvous by air- lift. From .Arabia to Australia—and beyond to New Zealand and the ice runways of Antarctica—the comings and goings of the airlift planes are today real scenes taken from Tunner’s vision. No one will disagree with Over the Hump that Tunner’s vision took a good many pokes in the eye over the years. Years when it was deemed necessary to deny one senior general officer the use of a transport to move his office equipment during the Korean War. (His priority was too low.) Years when it was necessary to appeal directly to Air Force Secretary Symington for engines, spare parts, and even mechanic’s wrenches to keep the Berlin Airlift going. (Let us draw the curtain of charity over the choleric effect this had on the theater commander who was charged with this support requirement.) Years when Tunner himself must have felt the effect of seeing his immediate successors at U.S. Air Forces in Europe and at mat s put on four stars to do a job he had been doing with only three. If Over the Hump should have been retitled The Battle of the Century, then who won? Certainly Tunner did. Four years after re- tirement his services have never been in greater demand. He is on the board of directors of the leading air freight airline operating over the Atlan- tic. He is a consultant to the Douglas Aircraft Company on their heavy air transport programs. His advice and counsel on airlift problems are sought regularly by senior military and civilian members of the Defense establishment. The high respect for his Opinions held by the Congress is reflected in the action of The Honorable L. Mendel Rivers in naming General Tunner “Mr. Airlift” Arriving at Ramey AFB to observe Operation Big Slam/Puerto Pine, Secretary of the Air Force Dudley following the Airlift Subcommittee hearings in Sharp is met by Military Air Transport Service 1960. commander, Lieutenant General William H. Tunner. But it is far from a one-sided victory. There Operation Big Slam was the last big MATS exercise are other winners also. before General Tunner’s retirement in 1960. 98 AIR UNIVERSITY REVIEW lift economics are in general use today. the headhunters below lick their chops and hone With the axiom of the nail and the horseshoe their knives. escalated into billion-dollar procurements for Or join hungry Hump flier Eddie Guilbert in global-ranging forces, it is unrealistic to gallop into clamoring for “eggis and flied lice” at the local battle dragging your depots behind. Chinese airstrip. And if you preposition what you need at point Or hoist a glass of Rhineland’s finest with X, how do you know that the battle won’t be fought Honorary Burgomeister Robert Hogg as he postu- at point Y? Or so goes Tunnel’s premise for which lates the bylaws of the Berlin Birdwatching and he fought again and again in the hot and cold war Airlift Operators’ Improvement Association. areas of the world. Or skim in over the “frozen Chosen” reservoir But the real winner is the reader of Over the with “Red” Forman in a bullet-holed C-47 to help Hump. He can join navigator Hal Sims in singing the embattled U.S. Marines fight their way out of a ring of Chinese bayonets. If 1 don’t hit Ascension Tunner knew his men and he knew his My wife will get a pension. mission. He demonstrates in Over the Hump that he Or he can churn over the Naga Hills with also knows how to tell an exciting tale of military ex-stunt pilot Andy Cannon in a balky C-46 while adventure. Long Beach, California Contributors

C olonel Jack W. Coleman ( D.B.A., Indiana Univer- sity) is Head, Systems Management Department. School of Engineering, Air Force Institute of Technology, and Associate Professor of Accounting. He is rated navigator and command pilot and served in the European Theater during World War IL.Recent assignments have been as Assistant to the Chief, Budget Division, Headquarters Air Materiel Command; Assistant Dean, School of Busi- ness. AF1T; Chief, Central Procurement Policies and Procedures, Headquarters Air Force Logistics Com- mand. His published works include "Efficiency and Working Capital Funds” in The Federal Accountant, March 1961.

M ajoh David C. D el l inger (Ph.D., Stanford Univer- sity' is Assistant Professor of Statistics and Operations Research, Systems Management Department, School of Engineering, Air Force Institute of Technology. He served as a fighter pilot during World War II and was recalled to active duty in 1951. He served as squadron and group operations officer, 471th Fighter Bomber Group, in Korea and in the United States. He studied industrial engineering at Stanford 1956-58; served on the Quality Control staff. Headquarters Air Materiel Area, from 1958 until his present assignment in 1961.

C aptain Douglas N. Jones (Ph.D., Ohio State University) is Associate Professor of Economics, Department of Economics and Geography, U.S. Air Force Academy. Previous service in- cludes a tour as Executive Officer to the Director of Transporta- tion, Hq Air Materiel Command, 1956-58, and consulting tours in the Office of the Assistant Secretary of Defense and in the Office of the Vice Chief of Staff, Operations Analysis.

Dn. H ans H. Z sc h ir n t (S c .D ., Technische Hochschule, Munich, Germany) is Senior Scientist, Data Sciences Laboratory. Air Force Cambridge Research Labora- tories. Massachusetts. During the war years he served as a physicist, later project leader, in the research es- tablishment of the Cerman Luftwaffe. In 1947 he came to the United States as part of "Operation Paper Clip” and joined the Watson Laboratories, Cambridge Field Station, predecessor of the present AFCRL, for work on microwave optics and ferro-resonnnt absorption of microwaves. His assignments at AFCRL have been with the Digital Communications Project; as Chief, Circuitry Group; Chief, Airborne Data Processing Project, 1954- 55; Chief, Data Processing Techniques Section. Com- puter Laboratory; and Assistant Chief, Computer and Mathematical Sciences Laboratory. Dr. Zschirnt has published numerous papers and reports in the fields of ultrasonics, radio* navigation, electronic countermeas- ures. and digital techniques. M ajor Robert L. A bl e (Ph.D., University of Kentucky) is C aptain Alan J. Grill (M.B.A., University of Chicago) is Associate Professor of Economics, Air Force Academy. He Chief, Officer Training, Webb AFB, Texas. He was com-J served with the Fifteenth Air Force in Italy during World missioned in 1953 upon completion of navigator training W ar II, Hying B -2 4 ’s. and in the Korean W ar flying B -26’s. and completed pilot training in 1956. He has served as I He has also served as a B—47 pilot in the Strategic Air Com- Armament Operations Officer, 92d Bomb Wing, SAC, Biggs! mand. Most of the rest of his career has been in education, AFB, Texas, 1953-55; Flight Commander, 465th Fighter including an AFROTC tour at Kentucky. He has published Interceptor Squadron, ADC, Griffiss AFB, New York, several articles on management development and related 1956-59; Flight Commander, 57th FIS, Keflavik, Iceland,] topics. At present Major Able is involved in activating the 1959-60; as student, AFIT, University of Chicago, 1960-61; Academy's new Management major and Master’s in Business and at Webb AFB since 1961, as Flight Instructor, Officer Administration program in cooperation with UCLA. Training Instructor, and in his present assignment since | 1963. Captain Grill is a graduate of the Academic Instructor Training, Randolph AFB, Texas, and of the Air Command and Staff College by correspondence, 1964.

C olonel John P. L is a c k (M.A., Ohio State University) is Technical Director, Air Force Educa- tional Requirements Board Secretariat, Maxwell AFB. After two years in the Illinois National Guard, he was commissioned and entered the Air Corps in 1936, receiving his regular commission in 1946. During World War II he served as Aircraft Maintenance Engineering Officer with heavy bombers in both the Pacific and European Theaters. Subsequent assignments have been with the Air Technical Service Command and the Continental Air Command; as Chief, Commitments and Requirements Branch, Operations and Training Division, Hq USAFE, 1951-55; as student, Armed Forces Staff College; and as Chief, Long Range Plans Branch, AFPDP, Hq USAF, for five and a half years preceding his current assignment.

C olonel Raymond L. T owns, USAF Retired, was for four and a half years Director of Information, Military Air Transport Service, preceding his retire- ment on 31 May 1964. He was a bomber pilot and operations officer in North Africa and Europe during World War II. Since then he has been closely as- sociated with airlift operations except for two years as special assistant to the USAF Surgeon General. He served as Flight Opetations Officer and later organized and directed operational production con- trol for the Berlin Airlift Task Force. During the Korean W ar he was Chief of Information, Far East Air Force’s Combat Cargo Command. In 1953 he became Chief of Public Information and Special Projects, Headquarters United States Air Forces in Europe. He is a graduate of the Air Command and Staff College, Air Force Special Staff School for Public Relations, and the Air Force Advanced Public Relations Course, Boston University. Colonel Towne is now Director of Public Relations. Douglas Aircraft Company, Long Beach, California. EDITORIAL STAFF

L ie u t e n a n t Colonel Ken n et h F. G antz, USAF, Ph.D., Editor J ack H. Mooney, M.A., Managing Editor L ieu t en a n t Colonel Harmon H. Harper, USAF, Chief, Acquisition Branch E dmund O. Barker, Financial and Administrative Manager J ohn A. W est c o t t , J r ., Art Director and Production Manager E nrique Gaston, Ph.D., Associate Editor, Foreign Language Editions W il l ia m J. D e P aola, Illustrator M ajor Robert G. S pa r k ma n , USAF, M.A., Editorial Project Officer

ADVISERS

C olonel Glen W. Clark, Hq Air Defense Command

C olonel W il l ia m J. M c G in t y , Hq Air Force Systems Command

C olonel Geo r g e Schexkein, Hq Tactical Air Command

L ie u t e n a n t Colonel Jack Rose, H

L ie u t e n a n t Colonel James F. S underman, Hq Pacific Air Forces

H arry A. H aberer, Hq Air Force Logistics Command

L a V er n e E. W oods, Hq Air Force Cambridge Research Laboratories

ATTENTION Air University Review is published to stimulate professional thought concerning aerospace doctrines, strategy, tactics, and related techniques. Its contents re- flect the opinions of its authors or the investigations and conclusions of its edi- tors and are not to be construed as carrying any official sanction of the Depart- ment of the Air Force or of Air University. Informed contributions are welcomed.