The Origins of Taxonomy Author(s): Peter H. Raven, Brent Berlin, Dennis E. Breedlove Source: Science, New Series, Vol. 174, No. 4015 (Dec. 17, 1971), pp. 1210-1213 Published by: American Association for the Advancement of Science Stable URL: http://www.jstor.org/stable/1732886 Accessed: 12/10/2009 12:41

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=aaas.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve and extend access to Science.

http://www.jstor.org speaking Mayans of southern Mexico, the Hanunoo of the Philippines, the Cantonese-speaking boat people of Hong Kong, the Guarani of Argentina, The of the Navajo, and many others. The prin- Origins Taxonomy ciples common to all these folk taxo- nomic systems, and to others that have A review of its historical shows been studied in sufficient depth, are as development why follows. taxonomy is unable to do what we expect of it. 1) In all languages, recognition is given to naturally occurring groupings of organisms. These groupings appear Peter H. Brent Dennis E. Breedlove Raven, Berlin, to be treated as psychologically discon- tinuous units in nature and are easily recognizable. They will be referred to here as taxa. By currenttaxonomic standards, there numbers of flies will never be seen by 2) These taxa are further grouped are probably about 10 million species a taxonomist, to say nothing of even into a small number of classes known of organismsin the world, of which we more obscure groups !such as the as taxonomic ethnobiologicalcategories. have in the past 218 years described, at Rickettsiae,which are of very high eco- These categories, definable in terms of some level, 10 to 15 percent. For more nomic importance. linguistic and taxonomic criteria, seem than 99 percent of the describedspecies, In this article, we shall attempt to to number five: unique beginner, life we know nothing more than a few mor- show why our present taxonomic sys- form, generic, specific, varietal. phological facts and one to several lo- tem is, in the face of the job for which 3) The five taxonomic ethnobiologi- calities where they occur. The human it has responsibility,inadequate. Being cal categories are arrangedhierarchical- population of the world, currently 3.7 basically a Renaissance codification of ly, and taxa assigned to each rank are billion, is growing at a rate which, if folk taxonomic principles made on the mutually exclusive. maintained, would lead to a doubling implicit assum,ptionthat the number of 4) The taxon found as a member of of the present size in 35 years. Yet it is organisms to be dealt with would per- the category unique beginner is often far from certain that the world can sup- haps be 25,000 to 50,000, it is incapa- not labeled linguisticallyby a single ex- port even present population levels in- ble of doing what we expect of it. Be- pression;that is, the most inclusive tax- definitely. Pollution on a world scale is fore our inventory of world organisms on, for example, plant, animal, is rarely increasingso rapidly that organismsare can be conducted in a way that is truly named. undoubtedly already becoming extinct meaningful and productive for scien- 5) Taxa that are members of the at a high rate. For example, tens or tific advance, we need to find new stan- category life form are invariablyfew in even hundredsof thousandsof kinds of dards for recording information about number, ranging from five to ten, and new synthetic molecules are being organismsin a readily retrievableform. these include a majority of all named dumped into the sea continuously; in taxa of lesser rank. almost all instances,their effects are un- 6) In most folk taxonomies,taxa that known. Developmentof Taxonomy are members of the category generic In view of the available taxonomic (3) are more numerous than life form manpowerand the enormousrate of ex- Man is by nature a classifying ani- taxa, but are nonetheless finite in num- tinction that will characterize the next mal. His continued existence depends ber, usually about 500. Some particular- century, it is doubtful that even 5 per- on his ability to recognize similarities ly aberrantgeneric taxa-for example, cent more of the world's organismscan and differences between objects and cacti, pineapple, cassowary, pangolin, be added to our inventory before the events in his physical universe and to platypus-or those that are of great eco- remaining 80 percent becomes extinct. make known these similarities and dif- nomic importance and interest may be Those familiar with tropical environ- ferences linguistically. Indeed, the very unaffiliated; that is, they are not in- ments, such as Paul Richards (1), pre- developmentof the human mind seems cluded in one of the life form taxa. dict that there will be no undisturbed to have been closely related to the per- 7) Specific and varietal taxa are, in tropical rain forest anywhere in the ception of discontinuitiesin nature. In general, less numerous than generics. world by the end of the present century. view of this, the study of folk taxo- Characteristically,they exist in sets of The tropical rain forest is the richest nomic systems, which has received a few members within a single generic. area in the world in species of plants great deal of interest in recent years, Sets of more than two members tend and animals, and by far the most poorly has a high significance in interpreting to refer to organisms of major cultural known. The vast majority of the esti- the logical processes going on in our importance, and sets of 20 or more mated million species of mites, of the own minds, as well as in understanding members inevitably do. Specific and quarter million to million species of the application and utility of the taxo- varietal taxa can be recognized linguis- ichneumonid wasps, and the unknown nomic systems themselves. tically in that they are commonly la- We beled in a binomial or trinomialformat Dr. Raven is director of the Missouri Botanical have (2) reviewed evidence sug- Garden and professor of biology at Washington gesting that there are a number of gen- that includes the name of the generic University, St. Louis. Dr. Berlin is associate pro- to which fessor of anthropology, at the University of eral characteristicscommon to the folk or specific they belong. California, Berkeley. Dr. Breedlove is assistant taxonomic systems that have been stud- 8) Intermediate taxa are those that curator of botany at the California Academy of one of the Sciences, San Francisco. ied in peoples as diverse as the Tzeltal- are immediatelyincluded in 1210 SCIENCE. VOL. 174 major life form taxa and that immedi- organisms included in the system are of Tournefort. Nevertheless, Linnaeus ately include taxa of generic rank. They known to the user. Given the contexts recognizedalmost no named taxa above are invariablyrare in natural folk tax- in which organisms are discussed and the generic level, even though he per- onomies, and, when evidence has been classified, there is generally nothing to ceived and discussed many such group- presented that unambiguously demon- be added to this knowledge by grouping ings. He presumably saw no need to strates their existence (4); they are not the organismsinto more inclusive cate- give them names, since he still consid- linguisticallylabeled. We have referred gories. Indeed, if there is a finite num- ered his genera to be limited in number to them as covert categories. ber of generic and other names for and memorizable. organisms that can be transmittedver- One interestingsimilarity between the bally in a language, intermediate cate- taxonomy of Linnaeus and the folk of Folk and Equivalence gories would have to be of very great taxonomic systems on which it was Early Published Taxonomies interestindeed to receive explicit labels. based has been pointed out by Walters (11). In folk taxonomies, organisms of Early biologists, such as Theophras- high cultural significance are apt to be tus, and later the herbalists merely LinnaeanTaxonomy and Its subdividedinto more generics, in terms wrote down folk taxonomic systems Predecessors of standard biological nomenclature, sharing completely the characteristics than are those of lower cultural sig- just enumerated. Both systems in- With the invention of movable type nificance (6). In just the same way, cluded relatively few specifics and va- in the mid-15th century, men began to Linnaeusrecognized many genera, each rietals in relation to the number of aspire to more and more comprehensive with few species, in plant families such As generics. early writers had no ready listings of the kinds of organisms. It as Brassicaceae and Apiaceae-plants means of duplicating and distributing began to seem worth while to attempt of high cultural utility that already had their works, they still confined them- to describe and name the many kinds Greek or Latin generics associatedwith selves to a limited number of generics of plants and animals being found in them. In contrast,he named few genera, in each domain. Generic plant names in this age of exploration, although the each with many species, in families such languages that have been fairly well biological dividends were at first mod- as Cyperaceaethat were of low cultural studied range from 250 to 800 forms, est. As travel to distant lands became utility and that had not been treated ex- regardlessof the richness of the envi- easier and easier, biologists began to tensively by his.predecessors. These his- ronment in which the people live. Ani- realize that the plants and animals of torical trends are still reflected in the mal names fall in the same range for different regions often differed greatly. level at which genera are recognized in generics (5). In all cases, including Thus, a herbal written by Dioscorides the respectivefamilies at present. early published taxonomic works, the in Greece was not apt to provide an In broad outlines, then, the system size of the basic set of generics appears adequate or useful description of the of Linnaeus was a codification of the to be controlled by the number of plants of central Germany some 15 folk taxonomy of a particular area of categories into which the known or- centuries later. The emphasis was still Europe, as has been stressedby Walters ganisms can be divided in a culturally on generic names, and, by the close of (12). It differed in principle from the significant fashion. We have shown the 17th century, Tournefort, in the unwrittenfolk taxonomy of the Tzeltal earlier that organisms of high cultural face of rapidly accumulating numbers speakers of Chiapas, Mexico, only in significance are apt to be subdivided of species, had stabilizedthe concept of the somewhatlarger numberof generics into more categories,in terms of West- the genus as central for biological tax- that was included and in the ern vastly biological nomenclature, than are onomy. He defined 698 genera (7), a greater number of specifics into which those of lower cultural significance numberstill consistent with the number they were divided. Both of these trends In an (6). operational system, espe- of generics recognizedin folk taxonom- were facilitated by the invention of one that is cially strictly verbal, the ic systems (5). Tournefortreviewed the movable type some 300 years earlier number of names cannot be multiplied genera that had been proposed and and made possible psychologically by beyond meaningful limits; many speak- tried to put them in a logical system (8). the expectation that the works would ers of the language must be famil- The situation was much the same at be widely distributedand useful. iar with each name that is passed down the time of Linnaeus;thus Cain (9) has But Linnaeus had to concern himself from to generation generationas part of pointed out that this 18th-centurytax- with relatively few kinds of plants and that As there language. are more and onomist founded his binomialsystem on animals. Had he even suspected the ex- more the names become less names, and the conception of a limited number of istence of the tens of thousandsof gen- less useful. memorizablegenera. Linnaeus felt that era to be discovered in the tropics and The lack of names for intermediate the genera must be separate, distinct other poorly known portions of the in folk and categories early written tax- units with distinct names, and natural world in the 19th and 20th centuries, onomies is probably related to these so that all would agree on their limits. he might not even have attempted the same influenceson language. For a giv- But Linnaeus had to deal with more synthesis of plant and animal classifica- en name to be assigned to a category genera than Tournefort. By 1737 he tion that he achieved so well. In an ef- two or more including generics, this in- included 935 in his Genera Plantarum fort to solve this difficulty,various post- termediate would have to category be (10), and in all of the editions of the Linnaeanauthors built up the curiously one that was often a subject of discus- work and its supplementshe presented deep and cumbersome taxonomic hier- sion and a point of reference. This is, diagnoses for 1336. The age of explo- archy that is characteristic of modern however, unlikely because, by the very ration had resulted in the discovery of classificationschemes. They added and nature of a folk taxonomy, all of the an uncomfortablylarge number of gen- named categories such as family, order, culturally significant properties of the era, as defined by the logical principles and phylum, and hoped that these at 17 DECEMBER1971 1211 least, being few in number, would be great deal. We know in fact that it has to focus on the process of classification more comprehensiblethan genera. not, but are still loathe to give up our rather than on the information being Some of the better marked and larger generalities,which we hope will help us gathered. families of flowering plants had been understandthe 107 kinds of organisms This general line of reasoning leads recognized as distinct units for many estimated to occur in the world. to important considerations for the centuries, but the first synoptical treat- In other words, the taxonomic system planning of future research. For ex- ment of the families of plants was pre- we use appearsto communicatea great ample, it is often argued that, if we sented by de Jussieu in 1789 (13). De deal about the organism being dis- know about the systematics of a par- Jussieu recognized exactly 100 families cussed, whereas in fact it communicates ticular group of organisms in detail, of plants, a number that could scarcely only a little. Since, in the vast majority we will be better able to utilize them in have resulted from anything but a con- of instances,only the describerhas seen biologicalcontrol programsand the like. scious effort to produce a strictly lim- the named organism,no one with whom This assumptionhas almost never been ited number of memorizableunits (11). he is communicatingshares his under- realizedin fact. What we have achieved He was trying to solve the dilemma to standingof it. The basis of communica- in biological control, pathology, and which the burgeoningnumbers of gen- tion present in folk taxonomic systems allied fields has been almost entirely era diagnosed by Linnaeus, who had dealing with a limited number of orga- the result of ad hoc studies of the prob- died in 1778, had given rise. Many of nisms is lacking, althoughwe still believe lems when they become of interest and his families were divided subsequently implicitly in its existence. Our system direct applicationof the results of these as more came to be known of the of names appears to achieve a reality studies. plants, and in 1920 Gundersen(14) was which it does not in fact possess. Taxonomic work has helped us only able to write casually, "The number of If we are truly interestedin recording to a limited extent in understandingthe families of vascular plants is generally information about organisms in a re- functioning of ecosystems, a problem considered as about 300." A few years trievable form, then we need a system that is of crucial importancefor human later, Hutchinson (15) pointed out that in which the points of reference are survival.Knowledge gained about a par- the delimitation of families is "very stable and the observationsare recorded ticular kind of organism outside of the much a matter of taste and personal in a standardformat. When a species is ecosystem being studied will often be idiosyncrasy"and to advocatethe recog- transferredfrom one genus to another, totally inaccurate for the area we are nition of about 400 for the flowering its name changes with serious loss to trying to understandand, thus, mislead- plants. How curiously these numbers the information retrieval capabilitiesof ing, even if it is available. At best, resemble the numbers of generics rec- the system as a whole. Yet such shifts equivalencies in the naming of orga- ognized in a very diverse sample of folk in generic position are one of a few nisms from different ecosystems may taxonomies (7)! mechanismswe have available to regis- provide a rough idea of what to expect, ter our changing appraisalof the rela- and no more. Clearly, if we are to tionshipsof a particularspecies. In fact, advance more rapidly in this vital field, Problemsfor Modern Taxonomy the current taxonomic system is hope- new taxonomic methodology seems to lessly inadequate as an information re- be needed. A folk taxonomic system is designed, trieval device, and it must be supplanted Completinga world survey of all or- not for information retrieval, but for with one allowing the characteristicsof ganisms is patently impossible; more- communicating about organisms with organisms to be handled and retrieved over, it has not been demonstratedto those who already understand the na- in a much more efficient manner. be of more than a very limited practi- ture of the organisms being discussed. cal, and of virtually no theoretical, im- These organisms and their culturally portance. It might be possible and de- significantfeatures are part of the active The Future sirable to complete the world survey of ethnobiologicalknowledge of most adult floweringplants, butterflies,or fleas;but speakers of the language. It would We have as yet named only about 15 it is clearly out of the questionfor many therefore be meaninglessto ask a Tzel- percent of the world's organisms and other groups and should not be taken tal speaker what propertiesthe taxa in- have no real chance of adding many to as a priori justification for the fund- dicated by his set of generic names have the total before the rest become extinct. ing of any proposed scientific work. in common. Only when the classifica- Despite this, we cling to the naive view When surveys of groups, particularly tory system is extended to hundreds of of Renaissance man and assume that tropical groups, are undertaken,the ap- thousands of poorly known organisms the extension of those folk principles propriate depth of the survey proposed do we begin to ask for a "definition"of deeply rooted in our collective psychol- should always be taken into account. genera and species. Confronted with ogy is the only appropriateway of deal- Perhaps blocking out families or genera this difficulty, the human mind is all ing with this diversity.We implicitly as- would be adequate for many groups; too ready to accept spurious general- sume that we know as much about a there may be no need whatever to de- ities such as the "biologicalspecies con- mite from the Amazon Basin as we do scribe and give names to the species. cept" or the earlier assumption,current about the mallard duck, and reflect this The "stockpiling"of specimens, espe- around the year 1900, that whole assumptionin our presumablyscientific cially from poorly known areas of the groups of related plants were merely but actually folk system of naming the world, continues to appearworth while, environmentalmodifications of one an- two kinds of organisms. We continue since, althoughwe cannot guess to what other. It would be comforting to find to conceal or lose most of our system- uses these specimens might be put in that evolution had produced a series of atic data when we make taxonomic de- 200 years, we can be certain that most identical units about which we know a cisions about organisms, since we tend of the species involved will be extinct. 1212 SCIENCE, VOL. 174 High priority should also be given (i) biology as well as a careful review of the difficulties of the system and ren- to taxonomic work that utilizes "unus- the principles upon which our taxo- der it still less useful for information ual" charactersor a broad spectrum of nomic system is based. retrieval. characters; (ii) to the accumulation of Folk taxonomies all over the world With modern electronic data process- information about organisms which are shallow hierarchicallyand comprise ing equipment, it has become possible does not seem to have direct taxonomic a strictly limited number of generic to record information about organisms, applicability;and (iii) to the search for taxa ranging from about 250 to 800 to retain this information in a data original ways of looking at the structure forms applied to plants and a similar bank, and to utilize it for various pur- of nature, including new methods of number applied to animals. These num- poses, including the construction of presenting "taxonomic" information. bers are consistent, regardless of the various taxonomic systems. The inven- One of the most significant trends in richness of the environment in which tion of high-speedelectronic data proc- modern systematics has been the de- the particular people live. Very few essing equipment is seen as analogous velopment of electronic data processing specific and varietal taxa are recog- to but more important than the inven- equipment. When information about nized in folk taxonomic systems. Un- tion of movable type in the history of organisms is entered into a data bank, til the invention of movable type in the systematicbiology. By using such equip- it can be summarized in any way that mid-15th century, written taxonomies ment to its full potentialities,we should might be desired, including the con- were simply records of the folk tax- be able to achieve a qualitative im- struction of taxonomic systems. The in- onomies of particular regions. Subse- provement in our perception of the formation that goes into the construc- quently, with the possibility for the living world. tion of these need not be wide distributionof it to systems lost, books, began References and Notes however, once the systems have been seem worth while to attempt to describe 1. P. W. Richards, The Tropical Rain Forest constructed; it is still available for re- and name all species of plants and ani- (Cambridge Univ. Press, 1952). trieval or recombinationin other ways. mals in the world. By the year 1700, 2. B. Berlin, D. E. Breedlove, P. H. Raven, Amer. Anthropol., in press. We can continue to have direct access 698 genera of plants were recognized; 3. H. H. Bartlett [Bull. Torrey Bot. Club 67, to the actual information that is avail- and by the year 1778, some 1350 gen- 349 (1940)] seems to have been the first to point out that the concept of the genus had able about a particular kind of orga- era, including tens of thousandsof spe- priority to that of the species in folk taxon- nism and need not generalizeso crudely cies. In 1789 de Jussieu interpolated omy, a position amply supported by subse- quent studies. from the fact that it happens to have the family as a higher level taxonomic 4. B. Berlin, D. E. Breedlove, P. H. Raven, been considered a species by one or category in an attempt to reduce the Amer. Anthropol. 70, 290 (1968). 5. Generic plant names for languages that have more taxonomists. Since the discontin- number of important units in the sys- been fairly well studied range from 250 to uities in the universe were tem to a memorablenumber. The fam- 800 forms; for example, Tzeltal 475, Hanun6o biological 822, Cantonese 267, Navajo 306, Seminole the great proving ground for the hu- ily is still the focal point in systems of 250, Hopi 350. Animal names fall in the man classification at same range for generic classes; for example, mind, which seems to have de- angiosperm present, Tzeltal 311, Hanun6o 461, Tzotzil 233, Nav- veloped largely as a device for dealing several hundred families being recog- ajo, insects only, 105, Maya 281 (see 2). 6. B. Berlin, D. E. Breedlove, P. H. Raven, with them effectively, we cannot pre- nized. Science 154, 273 (1966). dict what we may learn when we bring Problems with the taxonomic system 7. J. P. de Tournefort, Institutiones Rei Herba- riae (Paris, 1700). these discontinuitiesinto sharper focus stem largely from the fact that it is not 8. W. T. Steam, Introduction to facsimile re- with the aid of modern technology. designed as an informationretrieval de- print of C. Linnaeus, Genera Plantarum (1754). Historiae Naturalis Classica 3, v vice. In folk taxonomies, names are (1960). given to organisms and these are used 9. A. J. Cain, Proc. Linnean Soc. London 169, 144 (1958). Summary to communicate about the organisms 10. C. Linnaeus, Genera Plantarum (Uppsala, with others who already know .the cul- ed. 1, 1737). 11. S. M. Walters, New Phytol. 60, 74 (1961). There are approximately 10 million turally significant properties of the or- 12. --, Preslia 34, 207 (1962). kinds of in the world, of ganisms being discussed. In dealing 13. A. L. de Jussieu [Genera Plantarum (Paris, oiganisms 1789)] recognized exactly 100 families in the which we have described some 15 per- with the vast numbers of organisms first synoptical use of this category. cent. The of the human that we tend to 14. A. Gundersen, New Phytol. 19, 264 (1920). rapid growth exist, overemphasize 15. J. Hutchinson, The Families of Flowering population will cause most of the re- the process of classification and the Plants (Macmillan, New York, 1926), vol. 2. earth be- decisions it at of 16. We are grateful to the National Science mainderto disappearfrom the involves the expense Foundation for support through grants GS fore they are seen by a taxonomist. the information 'about the organisms 383, 1183, 2280, and GB 7949X. We thank Richard S. Cowan, Richard W. Holm, Roy These facts suggest a more rigorous that we are supposedly accumulating. L. Taylor, and S. M. Walters for their application of priorities in systematic Frequent changes in names exacerbate comments on this article.

17 DECEMBER 1971 1213