ENGLISH LINGUISTIC GUIDE

Despite the remarkable and irreversible changes

that have come up on the English language since the AngloSaxon p erio d

it has not yet reached a p oint of p erfection and stability

suchaswe sometimes asso ciate with Latin of the Golden Age

the language of Virgil Horace and others

DR ROBERT BURCHFIELD The English Language

in The Oxford Guide to the English Language

What a patchwork has b een our old saxon

by the bitter frost that nipp ed its early budding

and the constant habit of b orrowing thence resulting

the learned among usas well as the unlearned

though in very dierentwaysare constantly made to feel

The English language as wehaveitnow

is not so much a coherentgrowth as a disturb ed organism

PROF JOHN STUART BLACKIE Gaelic Self Taught

CONTENTS

ENGLISH ORTHOGRAPHY

Numb er of sp ellings

Sp elling number

Language Co de

Frequency of Sp elling

Sp elling

Diacritics

Reverse transcriptions

Sp elling columns

Transcriptions for corpus typ es

Transcriptions for English lemmas

Sp ellings for English headwords

Sp ellings for syllabied headwords

Transcriptions for wordforms

Sp ellings for plain wordforms

Sp ellings for syllabied wordforms

ENGLISH PHONOLOGY

Numb er of pronunciations

Pronunciation number

Status of pronunciation

Phonetic transcriptions

Computer phonetic character sets

Plain transcriptions

Syllabied transcriptions

Stressed and syllabied transcriptions

Example transcriptions

Phonetic patterns

ENGLISH MORPHOLOGY

Morphology of English lemmas

How to segment a stem

Typ es of analyses

The Derivation

The Comp ound

The Derivational Comp ound

The NeoClassical Comp ound

The NounVerbAx Comp ound

How to assign an analysis

The NounVerbAx Comp ound

Status and language co des

Derivationalcomp ositional information

Analysis typ e co des

Immediate segmentation

Complete segmentation at

Complete segmentation hierarchical

Other co des

Morphology of English wordforms

Inectional features

Typ e of ection

Inectional Transformation

ENGLISH SYNTAX

Word class co des letters or numbers

Sub classication Y or N

Sub classication nouns

Sub classication verbs

Sub classication adjectives

Sub classication adverbs

Sub classication numerals

Sub classication pronouns

Sub classication conjunctions

ENGLISH FREQUENCY

Frequency information for lemmas and wordforms

Frequency information from written and sp oken sources

Written corpus information

Sp oken corpus information

Frequency information for COBUILD corpus typ es

Frequency information for COBUILD written corpus typ es

Frequency information for COBUILD sp oken corpus typ es

ENGLISH ORTHOGRAPHY

Detailed and varied information is available on the ortho

graphic forms of headwords and wordforms You can cho ose

from a range of transcriptions they can b e syllabied or un

syllabied they can include or omit diacritics as explained

below or in some cases they come with the order of the

letters reversed or with the letters sorted alphab eticallyIn

addition there are columns whichtellyou the number of

letters or syllables a particular transcription contains

This flex window is the menuyou see for a lemma or a

wordform lexicon when you cho ose the Orthography option

of the rst ADD COLUMNS menu

ADD COLUMNS

Number of spellings

Spelling number N

Language code

Frequency of spelling

Spelling

TOP MENU

PREVIOUS MENU

NUMBER OF SPELLINGS

This option in the ADD COLUMNS menu is a column which

tells you how manyways eachlemmawordform or abbrevi

ation according to the typ e of lexicon you are using can b e

sp elt For the verb lemma generalize this column has the

value which means there are twopossibleways of sp elling

it Unless you construct a restriction on your lexicon gener

alize will o ccur twice one row using the form generalizethe

other using the form generalise

This column is particularly useful when you wanttoidentify

words whichhave sp elling variants To exclude from your

english linguistic guide

lexicon all items whichonlyhave one p ossible sp elling con

taining instead those whichcanbespeltina number of ways

you can construct an expression restriction which simply

states that the numb er of sp ellings must b e greater than

OrthoCnt

The flex name and description of this column are as follows

OrthoCnt Number of spellings

OrthoCntLemma

SPELLING NUMBER

Just as the very rst available ADD COLUMNS option is a

number which uniquely identies each lemma or wordform

according to the typ e of lexicon you are using so this

column uniquely identies every sp elling to b e found for each

lemma or wordform

If you are using a lemma lexicon the sp elling variants are

given in the form of headwords with or without syllable

markers For example the verb generalize has two sp ellings

one is the form generalize and another is the form generalise

These have the sp elling numb ers and resp ectivelyIfyou

use a syllabied stem representation in place of the plain

headword representation sp elling takes the form general

ize and sp elling the form generalise

This means you can use the universal sequence number to

identify a particular lemma or wordform dep ending on the

typ e of lexicon you are using and then the sp elling number

to identify the dierent individual sp ellings used for each

lemma Usually no preference is indicated by the sp elling

numb ers eachvariant has generalize is as valid as gener

aliseHowever the numb er sp elling is always an acceptable

British from and any American variants are always given a

higher numb er Thus the lemma monologue has the British

form monologue as its numb er sp elling and the American

form monolog as its numb er sp elling You can nd out

whether a sp elling is British or American by using the Or

thoStatus column describ ed b elow

One imp ortant p oint to rememb er is that the sp elling number

can b e used to eliminate unwanted rows from your lexicon

If you only want to see one sp elling for each lemma or

Sp elling number

wordform you should construct a restriction which states

that only rows with a sp elling numb er equal to are to b e

included in the form OrthoNum If you dont do this

you usually end up with lexicons that are to o long b ecause

they needlessly rep eat certain pieces of information Takethe

example generalize again only this time imagine you wantto

know its pronunciation rather than the various ways it can

b e sp elt You create a lexicon with three columns one giving

the sp elling numb er one giving the orthography of the stem

and one giving the pronunciation of the headword Without

the restriction flex returns tworows for generalize

Sp elling Number Headword Pronunciation

generalize dZEnrl aIz

generalise dZEnrl aIz

generalize dZEnrla Iz

generalise dZEnrla Iz

This is unnecessary since you are interested only in the

p ossible pronunciations The extra row merely gives you a

sp elling variant while the pronunciations remain the same

When you include the restriction OrthoNum however

only the rows with the numb er sp elling are included

Sp elling Number Headword Pronunciation

generalize dZEnrl aIz

generalize dZEnrla Iz

And of course the more lemmas your lexicon contains the

greater the numberofeliminated lines b ecomes simply as a

consequence of adding this imp ortant restriction

If you are particularly interested in sp elling variation then

do not add this OrthoNum restriction that wayyou

get to see all the variant orthographic forms of each lemma

Otherwise whenever you just want to use a simple ortho

graphic transcription as a means of representing the lemma

in your lexicon always rememb er to insert it

The flex name and description of this column are as follows

OrthoNum Spelling number OrthoNumLemma

english linguistic guide

LANGUAGE CODE

For every dierent sp elling there is a co de which tells you

whether it is an acceptable British form B or whether it is

only ever an American form A This applies to lemmas and

wordforms A British sp elling is always the rst one given

that is its sp elling numb er is always while one which

only o ccurs in American English is never the rst form and

always has a sp elling numb er greater than

Sp elling Status Example

typ e co de

British B adviser

American A advisor

Table Orthographic status co des for English sp ellings

The flex name and description of this column are as follows

OrthoStatus Status of spelling

OrthoStatusLemma

FREQUENCY OF SPELLING

There are gures available which tell you how frequently each

sp elling of each lemma or wordform o ccurs in the cobuild

corpus along with deviation gures whichgive a range of

error for each frequency They dier from the main frequency

gures in that they are sp ecic to one sp elling whereas the

frequency columns prop er refer to the more general frequency

counts for the whole lemma or for eachwordform

To arrive at these gures a count has to b e made of the

numb er of times each string o ccurs in the current million

word version of the cobuild corpus This lets you see that

the string b eauty for example o ccurs times and that

truth o ccurs times and these gures are the sp elling

frequencies for the wordforms which are sp elt that way

Usually but not always this string frequency is the same as

the wordform frequency and its then p ossible to formulate

sp elling frequencies for each lemma This simply means

adding together the frequencies for eachwordform in each

inectional paradigm Thus the sp elling frequency for the

Frequency of Sp elling

lemma b eauty is the sp elling frequency for the wordform

b eauties plus the sp elling frequency for the wordform

b eauty giving a total of Similarly the sp elling fre

quency for the lemma truth is truths plus truth

giving a total of

On the o ccasions when the string frequency cannot b e linked

to just one lemma an alternative plan of action is used Take

as an example the sp elling tender which can refer to four

dierent lemmas the rst is the adjective meaning soft or

gentle the second is the verb meaning oer or presentthe

third is a noun meaning nancial estimate or prop osition

and the fourth meaning the wagon which comes b ehind a

steam engine The problem is that the string frequency can

b e assigned to any of these lemmas it is always ambiguous

Toovercome this problem it is p ossible to checkevery o ccur

rence of tender in the corpus and work out exactly howmany

b elong to each of the four lemmas To a certain extent this

can b e done by computer program but celex underto ok the

task by hand reading o ccurrences in context and then decid

ing to which lemma the ambiguous string b elongs This ap

proach clearly requires more time but the investment yields

amuch more dep endable result The problem is though

that the words which require disambiguationand there are

approximately of themare usually very frequent

Disambiguating all the o ccurrences of just one word could

involve reading thousands of corpus sentences Toavoid this

a random sample of o ccurrences is taken from the corpus up

to a maximum of whenever the frequency is greater than

Disambiguating such a set pro duces a simple ratio

which can b e used to calculate the nal frequency gure

The string tender o ccurs times and after examining

o ccurrences of the word in the cobuild corpus out of the

hundred were gentle were the verb oer was the noun

nancial prop ositionandwas the noun wagon The ratio of

one meaning to the other is thus The

frequency of tender that is the adjective gentle is then

multipliedby which is while the frequency of the

verb oer ismultipliedby which is The noun

nancial prop osition and the noun wagon share the frequency

of multiplied by which rounded up comes to four

However the story is still not complete Occasionally it is im

p ossible to decide which lemma a particular sp elling b elongs

english linguistic guide

to The contraction Ill is one suchword since it can mean

I will or I shallInsuch cases a safe unambiguous decision

cannot b e made and the ratio is said to b e If there

are three p ossible options the ratio is

andsoon

The result of this work is that you have an accurate fre

quency gure for each sp elling of eachlemmawordform or

abbreviation in the database and that gure is contained in

this column the flex name and description of whichisas

follows

CobSp ellFreq Spelling frequency COBUILD m word corpus

CobSp ellFreqLemma

How accurate are the gures in the CobSp ellFreq column

The answer is that if there are no ambiguities to b e resolved

then the gures are naturally completely accurate This

is true for most of the words in the database From the

ab ove description though its clear that in certain cases a

degree of approximation is included When ambiguities do

o ccur then it is p ossible to calculate a deviation gure which

sp ecies the range of error to an accuracy of at least

This is the required formula

r

p
p N
n

N

n N

where N is the frequency of the word as a whole n is the total

number of words whichwere disambiguated in the random

sample and p is the ratio gure for the word when it b elongs

to one particular lemma Thus for tender the adjective

gentle N is n is and p is and the formula

gives as the deviation This means that the true frequency

for this form of tender is almost certain certain at

leastto lie b etween and

Occasionally you may come across cases where the deviation

gure is greater than or equal to the frequency gure itself

This indicates that you are dealing with a sp elling which

cannot b e disambiguated as with the example Ill discussed

ab ove While the frequency gures in such cases are arbi

trary the accompanying deviation gures are accurate

So while CobSp ellFreq gives the disambiguated frequency

gure for each sp elling this column indicates the statistical

Frequency of Sp elling

deviation of that gure Its flex name and description are

as follows

CobSp ellDev confidence deviation COBUILD m word

CobSp ellDevLemma corpus

Finally rememb er that the columns describ ed here refer only

to the frequencies of individual sp ellings most of the fre

quency information is dealt with in section English Fre

quency

SPELLING

Before dening the sp ecic sp elling columns available with

b oth of the English lexicon typ es its worth considering

a few imp ortant general features which apply to manyof

the columns namely diacritics and reversed transcriptions

After that come the individual sp elling columns themselves

DIACRITICS

As you work your waydown the ADD COLUMN menus you

can see that on several o ccasions the last menu in the series

allows you to select transcriptions whichcontainor omit

diacritics Diacriticsarethe accents written ab ove certain

characters as a guide to pronunciation Usuallyonlyforeign

words use such markers consistently in English words like

vicuna or soupcon or debacle These sp ecial accented char

acters are eightbit characters designed for use on certain

digital terminals the vt and newer terminals If you

use such a terminal or can get your own terminal to emulate

it then you lo ok at the diacritics columns with no problems

at all If you have a completely dierent terminal you can

still use diacritics columns by selecting the MODIFY COLUMNS

option CONVERT to change the digital eightbit co des to

the form your terminal needs to pro duce the same diacritic

characters

To do this you need a table of the digital eightbit co des

that celex uses such as the one given in part of the

manual the App endicesInityou can nd out the hexa

decimal co des of the letters you need to convert You also

need a table of the co des your terminal uses to pro duce the

english linguistic guide

same diacritical markers The example that follows converts

all the digital eightbit co des that are used in the English

database to their msdos equivalents as dened in the

olivetti msdos User Guide The characters whichoccur

with diacritic markers are as follows a a c e e e o

n andu When you reachthe MODIFY CONVERSION win

dow rst select a column whichcontains transcriptions with

diacritics then typ e in the following string

xxF

xExxEx xExxExA

xExxEAx xEFxBxFx

xFxAxFCx

Once installed this pattern will convert all the diacritic char

acters whenever you SHOW or EXPORT the column If youre

new to the pattern matcher and its capabilitiesthenitmay

app ear very mysterious but in fact its straightforward Read

the next couple of paragraphs for a full explanation

The rst line indicates that one or more normal asci i co des

those with hexadecimal values b etween and F are al

lowed

The remaining lines indicate the changes that must b e made

to any bit characters that o ccur The pattern matcher uses

the sign to indicate a conversion the element to the left

of the is converted to the element on the right This use

of the sign is dierent from the wildcard function it has

in an expression restriction or query The pattern matcher

also uses the symb ols x to mean that the twocharacters

which follow form a hexadecimal co de thus in the digital

eightbit co de xF actually means nInthemsdos co ding

set the same n character is represented bythecode xA

So to tell the pattern matcher to convert from a digital n

to an msdos nyou must typ e xFxA

So far this accounts for one diacritic character Toconvert

all the diacritic characters you have to add extra parts to

the pattern as appropriate until you end up with a pattern

like the one ab ove Each element is separated bytheor

marker The whole pattern comes b etween brackets fol

lowed by an asterisk at the end which means the

word may b e made up of zero or more of the elements b etween

the brackets

Reverse transcriptio ns

REVERSE TRANSCRIPTIONS

Transcriptions without diacritics are often available in re

verse order eachitemisgiven back to front Thus back

is given as kcab The reason for this is that with a draft

lexicon lo oking up word endings can b e done much more

quickly when you use reverse transcriptions

SPELLING COLUMNS

This section sets out the columns with sp ellings available for

each lexicon typ e First there is a short subsection dealing

with corpus typ e transcriptions then a longer subsection on

the headword transcriptions available with a lemma lexicon

nishing up with a subsection on wordform transcriptions

TRANSCRIPTIONSFOR CORPUS TYPES

One column is available It gives plain transcriptions which

include lower case letters hyphens full stops ap ostrophes

round brackets and digits If youre not sure exactly what

corpus typ es are check part of the manual the Intro duc

tion to nd out The flex name and description of this

column are as follows

Typ e Graphemic transcription

TRANSCRIPTIONSFOR ENGLISH LEMMAS

The English lemma is always represented by the headword

as describ ed in the Intro duction section When you

cho ose a column whichcontains orthographic transcriptions

of headwordsitisasifyou are cho osing the b oldtyp e head

word in a dictionary All the other columns in the database

contain information sp ecic to individual headwords so the

main function of the orthographic transcription is to identify

any other information you lo ok up lo oking at a list of

lemma frequency gures isnt meaningful unless you can see

the lemmas they refer to However you may not always

need to see the orthographic form of the headword if youre

lo oking for phonetic transcriptions with certain interesting

syllablenal characteristics sayyou may not b e interested

in the orthographic headword in whichcaseyou neednt

english linguistic guide

keep it on viewandyou mighteven want to miss it out of

your lexicon altogether

Describ ed b elowareseveral dierent forms of orthographic

transcription and each form is assigned its own column The

rst distinction you can makebetween them is whether or

not syllable makers are included Thereafter you can cho ose

between backtofront transcriptions transcriptions with or

without diacritics transcriptions which consist only of lower

case characters and even transcriptions with the letters of

the headword reordered alphab etically Read the details

and then cho ose the columns whichbesthelpyou to build

useful lexicons

SPELLINGS FOR ENGLISH HEADWORDS

There are six columns available which do not giveany indi

cation of the orthographic syllabication they just deal with

the letters in each headword

ADD COLUMNS

Without diacritics

Without diacritics reversed

With diacritics

Purely lowercase alphabetical

Purely lowercase alphabetical sorted

Number of letters

TOP MENU

PREVIOUS MENU

The rst column has information which is basic to the other

ve columns It simply contains headwords comp osed of

upp er and lower case characters hyphens and ap ostrophes

with no diacritics or any other alterations So the headword

which represents the verbal family of inections walk walks

walking and walked is quite simply walkFor information

ab out which forms are used as headwords see the Intro

duction section The flex name and description of this

column is as follows

Head Headword without diacritics HeadLemma

Sp ellings for English headwords

The second column contains the same transcriptions as the

rst only the order of the letters is reversed Thus the

headword walk is given as klawandincrease is given as

esaercni However with words like repap er you might not

notice much of a dierence And pro ceed cautiously with the

word embargo thereverse transcription is ograbme The

flex name and description of this column are as follows

HeadRev Headword reversed

HeadRevLemma

The third column gives sp ellings which include diacritics as

well as the basic upp er and lower case characters hyphens

and ap ostrophes of the basic transcriptions So while the

rst column gives the plain form cloissone this column in

cludes the authentic French acute accent cloissoneLikewise

debacle b ecomes debacle and deshabille b ecomes deshabille

The characteristics of diacritics are describ ed in section

ab ove The flex name and description of this column are

as follows

HeadDia Headword diacritics

HeadDiaLemma

The fourth column contains the same basic transcription as

the rst except that any upp er case letters which o ccur are

reduced to lower case and any nonalphab etic characters

are removed Thus Jeremiah b ecomes jeremiah and Sax

b ecomes sax

Such a column is useful when youre trying to sort a list of

words into true alphab etical order as opp osed to asci i order

Each letter whether upp er or lower case has a dierent

asci i numb er and computers usually sort and order letters

on the basis of these numbers Because lower case letters

all have higher numb ers than upp er case ones the results

of a sort program arent always what you exp ect However

with this column since all the characters are lower case the

problem do esnt arise So to make a le which contains

a true alphab etical list of plain headwords make a lexicon

which consists of the plain headwords column Head and

this column and when you export it put a against

this purely lower case column and in this way alphab etic

english linguistic guide

normality will b e restored The flex name and description

of this column are as follows

HeadLow Headword lowercase alphabetical

HeadLowLemma

The most imp ortant feature of the fth and last column is

that the letters whichmake up each headword are sorted into

alphab etical order This do es not refer to the dictionarylike

alphab etical order of headwords listed in the database its

to do with the letters within eachword And in addition

any upp er case characters are reduced to lower case and

nonalphab etic characters are removed Thus for example

Jeremiah b ecomes aeehijmr and dread p erhaps confusingly

b ecomes adder Using this column anagrams can b e solved

quickly and searches for words containing certain numbers of

letters can b e carried out with ease creating a query which

lo oks for aaa in this column can return a list of words from

another column whichcontain at least three a characters

The flex name and description of this column are as follows

HeadLowSort Headword lowercase alphabetical sorted

HeadLowSortLemma

The sixth and last column contains counts of the number

of letters in each headword Here letters means any upp er

or lower case alphab etic characters excluding hyphens and

ap ostrophes This means that sometimes the length of a

word is dierentfromthenumb er of letters it contains the

numb er of letters in focsle for example is The flex name

and description of this column are as follows

HeadCnt Headword number of letters

HeadCntLemma

SPELLINGS FOR SYLLABIFIED HEADWORDS

There are two columns whichcontain headwords with their

orthographic syllable markers In these columns a hyphen

marks the b oundary b etween each pair of syllables within the

headword Thus the plain headword abandonment is given as

abandonment There is a third column relating to syllabi

ed headwords and it tells you the numb er of orthographic

syllables eachheadword has

Sp ellings for syllabied headwords

ADD COLUMNS

Without diacritics

With diacritics

Number of syllables

TOP MENU

PREVIOUS MENU

The rst column contains the basic headwords plus syllable

markers each transcription consisting of upp er and lower

case characters hyphens and ap ostrophes The flex name

and description of this column are as follows

HeadSyl Headword syllabified

HeadSylLemma

The second column contains the same headwords as the rst

except that diacritics are included where appropriate The

flex name and description of this column are as follows

Headword syllabified diacritics HeadSylDia

HeadSylDiaLemma

Some p eople like to use only partially syllabied headwords

that is syllabied transcriptions which omit the rst syl

lable marker if the rst syllable consists of only one let

ter For example the partially syllabied transcription of

abandonment would b e abandonmentSuch transcriptions

are useful for automatic hyphenation programs since typ o

graphic convention says that a word divided at the end of

a line should consist of more than one character To obtain

transcriptions in this form you can use the CONVERT option

of the MODIFY COLUMNS menu When you reachthe MODIFY

CONVERSION window select a column containing normal syl

labied headwords and then typ e the following string

This means rst there is one character of some sort Then

if there is a hyphen followed byacharacter whichis not

english linguistic guide

ahyphen convert the hyphen into nothing then there are

zero or more other characters of some sort Thus whenever

you SHOW or EXPORT your lexicon the syllabied transcrip

tions will always app ear in partially syllabied form Two

hyphens together after a rst letter indicate that there is an

orthographic hyphen as opp osed to a syllable marker in the

sp elling at this p ointasinTshirt for example They are

left as twohyphens to dierentiate this sort of hyphen from

the other syllable markers the word mightcontain

The third and last column for syllabied headwords tells you

how many syllables eachheadword contains The number of

syllables in the word abandonment for example is The

flex name and description of this column are as follows

HeadSylCnt Number of orthographic syllables

HeadSylCntLemma

TRANSCRIPTIONSFOR WORDFORMS

Wordforms are the words we use in everydayspeechand

writing Elsewhere in the database families of wordforms

inectional paradigms are represented by one form the

lemma When you work with a wordform lexicon all the

wordforms are available as separate entries not just one

representative form All the other columns in the database

contain information sp ecic to individual wordforms so the

main function of the orthographic transcription is to identify

any other information you lo ok up lo oking at a list of

syntactic class co des isnt very meaningful unless you can see

the wordforms they refer to A full description of the prop

erties of wordforms can b e found in part one of the manual

the Intro duction under the section called Lexicon typ es

Orthographic transcriptions of wordforms are available either

with or without syllable markers The next section deals

with plain unsyllabied transcriptions and the one after

that deals with syllabied transcriptions

SPELLINGS FOR PLAIN WORDFORMS

Describ ed b elowareseveral dierent forms of orthographic

transcription and each form is assigned its own column

using the ADD COLUMNS menushown b elow you can cho ose

between backtofront transcriptions transcriptions with or

Sp ellings for plain wordforms

without diacritics transcriptions which consist only of lower

case characters and even transcriptions with the letters of

eachwordform reordered alphab etically Read the details

below and then cho ose the columns which b est help you to

build useful lexicons

ADD COLUMNS

Without diacritics

Without diacritics reversed

With diacritics

Purely lowercase alphabetical

Purely lowercase alphabetical sorted

Number of letters

TOP MENU

PREVIOUS MENU

The rst column contains information which is basic to the

other ve columns It simply contains wordforms comp osed

of upp er and lower case characters hyphens and ap ostrophes

with no diacritics or any other alterations

Word Word

The second column contains all the wordforms to b e found

in the rst column except that the order of the letters is

reversed Thus the wordform walks is given as sklawand

increased is given as desaercni However with wordforms

like deied you might not notice a big dierence And pro

ceed cautiously with the word desserts b ecause in this col

umn it has to b e stressed The flex name and description

of this column are as follows

WordRev Word reversed

The third column gives sp ellings which include diacritics as

well as the basic upp er and lower case characters hyphens

and ap ostrophes of the basic transcriptions The character

istics of diacritics are describ ed in section ab ove The

flex name and description of this column are as follows

WordDia Word diacritics

english linguistic guide

The fourth column contains the same basic transcription as

the rst except that any upp er case letters which o ccur are

reduced to lower case Thus Peking becomes p eking and

Uranus becomes uranus In addition any nonalphab etic

characters hyphens ap ostrophes are removed Such a col

umn is useful when youre trying to sort a list of words into

true alphab etical order as opp osed to asci i order Each

letter whether upp er or lower case has a dierent asci i

numb er and computers usually sort and order letters on

the basis of these numbers Because lower case letters all

havehighernumb ers than upp er case ones the results of

a sort program arent always what you exp ect However

with this column since all the characters are lower case the

problem do esnt arise So to make a le which contains

a true alphab etical list of plain wordforms make a lexicon

which consists of the plain wordforms column Word and

this column and when you export it put a against

this purely lower case column and in this way alphab etic

normality will b e restored The flex name and description

of this column are as follows

WordLow Word lowercase alphabetical

The most imp ortant feature of the fth and last column is

that the letters whichmakeupeachwordform are sorted into

alphab etical order This do es not refer to the dictionarylike

alphab etical order of wordforms listed in the database its

to do with the letters within eachword And in addition

any upp er case characters are reduced to lower case Thus

Peking is given as egiknpandUranus as anrsuu Another

example is the word editorials whose sorted form is adei

ilorstInterestingly enough adeiilorst is also the sorted form

of the word idolatries Using this column anagrams can

b e solved quickly and searches for words containing certain

numb ers of letters can b e carried out with ease creating a

query which lo oks for aaa in this column can return a list

of words from another column whichcontain at least three

a characters The flex name and description of this column

are as follows

WordLowSort Word lowercase alphabetical sorted

Sp ellings for plain wordforms

The sixth and last column contains counts of the number

of letters in eachwordform Here letters means any upp er

or lower case alphab etic characters excluding hyphens and

ap ostrophes This means that sometimes the length of a

word is dierentfromthenumb er of letters it contains the

numb er of letters in shouldnt for example is The flex

name and description of this column are as follows

WordCnt Word number of letters

SPELLINGS FOR SYLLABIFIED WORDFORMS

There are two columns whichcontain wordforms with their

orthographic syllable markers In these columns a hyphen

marks the b oundary b etween each pair of syllables within the

wordform Thus the plain wordform abandoning is given as

abandoning There is a third column relating to syllabi

ed wordforms and it tells you the numb er of orthographic

syllables eachwordform has

ADD COLUMNS

Without diacritics

With diacritics

Number of syllables

TOP MENU

PREVIOUS MENU

The rst column contains wordforms plus syllable markers

Each transcription consisting of upp er and lower case char

acters hyphens and ap ostrophes The flex name and de

scription of this column are as follows

WordSyl Word syllabified

The second column contains the same wordforms as the rst

except that diacritics as explained in section are in

cluded where appropriate The flex name and description

of this column are as follows

WordSylDia Word syllabified with diacritics

english linguistic guide

Some p eople like to use only partially syllabied headwords

that is syllabied transcriptions which omit the rst syl

lable marker if the rst syllable consists of only one letter

For example the partially syllabied transcription of aban

doning is abandoning Such transcriptions are useful for

automatic hyphenation programs since typ ographic conven

tion says that a word divided at the end of a line should

consist of more than one character To obtain transcriptions

in this form you can use the CONVERT option of the MODIFY

COLUMNS menu When you reachthe MODIFY CONVERSION

window select a column containing normal syllabied word

forms and then typ e the following string

partpartpartpart

This basically means call the rst character by the name

part The second character mayormay not b e a hyphen

Any subsequentcharacters are called part Rewrite the

whole word as part plus part Only the parts of the word

assigned to part or part are rewritten thus excluding the

rst hyphen whenever it o ccurs b ecause it is not assigned

to anyvariable When you SHOW or EXPORT your lexicon

the syllabied transcriptions will always app ear in partially

syllabied form However note that on this o ccasion when

a double orthographic hyphen o ccurs after the rst letter

only one of the twohyphens is written So if you ever do see

ahyphen as the second letter in your converted column you

know for sure that it is actually an orthographic and not a

syllabic hyphen

The third and last column for syllabied wordforms tells you

howmany syllables eachwordform contains The number of

syllables in the word abandoning for example is The

flex name and description of this column are as follows

WordSylCnt Word number of orthographic syllables

English phonology

ENGLISH PHONOLOGY

Phonetic transcriptions are available for each lemma and

wordform in the database They are sp ecied in four dif

ferentcharacter sets and variantpronunciations are also

given whenever they o ccur The transcriptions you cho ose

can also include syllable markers or stress markers Each

pronunciation has a unique identication numb er so that in

conjunction with the lemma or wordform numb er you can

identify every single pronunciation in the database Also

available are cv patterns stress patterns and phoneme

and phonetic syllable counts In addition when you are

using a wordform lexicon you can get phonetic information

and other information to o ab out the lemmas of anyofthe

wordforms The sections b elow deal with the information

under the headings which corresp ond to those used in the

ADD COLUMNS menus

ADD COLUMNS

Number of pronunciations

Pronunciation number N

Status of pronunciation

Pronunciation

Phonetic Patterns

TOP MENU

PREVIOUS MENU

Exactly the same columns are available for lemma lexicons

and wordform lexicons so all the phonetic column descrip

tions and denitions are equally valid for b oth lexicon typ es

NUMBER OF PRONUNCIATIONS

This option in the ADD COLUMNS menu is a column which

tells you how manyways each lemma or wordform according

to the typ e of lexicon you are using can b e pronounced For

the lemma dexterous this column has the value which

english linguistic guide

means there are two p ossible ways of pronouncing it Unless

you construct a restriction on your lexicon dexterous oc

curs twice one rowwith dEkstrs and another

row with dEkstrs these examples are from the

PhonSAM column

This column is particularly useful when you wanttoidentify

words whichhavepronunciation variants To exclude from

your lexicon all lemmas or wordforms which only have one

p ossible pronunciation containing instead those which can

b e pronounced in a number of ways you can construct an

expression restriction which simply states that the number

of pronunciations must b e greater than PhonCnt

The flex name and description of this column are as follows

Number of Pronunciations PronCnt

PronCntLemma

PRONUNCIATION NUMBER

Just as the very rst available ADD COLUMNS option is a

number whichuniquelyidenties each lemma or wordform

according to the typ e of lexicon you are using so this

column uniquely identies every pronunciation to b e found

for each lemma wordform or abbreviation

For example the noun dexterous has twopronunciations

rst dEkstrs and second an alternativeform

dEkstrs These have the pronunciation num

b ers and resp ectivelyIfyou use a syllabied transcrip

tion in place of the plain transcription pronunciation takes

the form dEkstrs and pronunciation the form dEk

strs

This means you can use the universal sequence number to

identify a particular lemma or wordform and then the pro

nunciation numbertoidentify the dierent individual pro

nunciations given for each lemma Moreover the pronun

ciation numb er allows you to identify quickly the primary

pronunciation as laid down in the English Pronouncing Dic

tionary by Daniel Jones AC Gimson and Susan Ramsaran

b ecause such forms are always rst in the list for every

lemma or wordform the numb er pronunciation is the pri

mary form The classications of variant pronunciations

Pronunciation number

are dealt with fully in the next section on the status of

pronunciation column

One imp ortant p ointtorememb er is that the pronunciation

numb er can b e used to eliminate unwanted rows from your

lexicon If you only want to see one pronunciation for each

lemma or whatever you should make a restriction which

states that only rows with a pronunciation numb er equal

to are to b e included in the form PronNum If you

dontdothisyou usually end up with lexicons that are

to o long b ecause they needlessly rep eat certain pieces of

information Take the example dexterous again in a lexicon

with three columns one giving the pronunciation numb er

one giving the orthography of the headword and one giving

the pronunciation of the headword Without the restriction

flex returns tworows for dexterous

Pronunciation Headword Pronunciation

Number

dexterous dEkstrs

dexterous dEkstrs

You may nd this unnecessary needing to see only one de

fault pronunciation the extra row merely gives you an extra

pronunciation When you include the restriction PronNum

however only the row with the preferred sp elling is

included

Pronunciation Headword Pronunciation

Number

dexterous dEkstrs

And of course the more lemmas or wordforms your lexicon

contains the greater the numb er of eliminated lines b ecomes

simply as a consequence of adding this imp ortant restriction

For some words there are manyvariants given transitional

has fortypossiblepronunciations all of which are given on a

separate row in the database If you are particularly inter

ested in pronunciation variation of this kind then do not add

the PronNum restriction that wayyou get to see all the

variant forms Otherwise whenever you just want to use a

simple phonetic transcription always rememb er to insert it

english linguistic guide

The flex name and description of this column are as follows

PronNum Pronunciation ID number

PronNumLemma

STATUS OF PRONUNCIATION

For every dierentpronunciation there is a co de which tells

you whether it is a primary pronunciation P or a sec

ondary pronunciation S This applies to b oth lemmas and

wordforms According to the English Pronouncing Diction

ary primary and secondary forms are all standard forms but

primary forms are heard more frequently

The rst pronunciation given for eachword that is the

pronunciation with PronNum is usually a citation form

the sort of pronunciation you are most likely to hear if you

asked a sp eaker of standard English to pronounce a par

ticular word by itself It is always given the status co de P

for primary pronunciation For example the numb er one

pronunciation of transparent is trnsprnt

Sometimes stylistic variants of this rst form are recorded

variants which indicate the highly frequent elision of sounds

that o ccur in connected sp eech Since the formal pronun

ciation of a word is quite rare the stylistic variants are

probably heard more often than the citation form The

second variantfor transparent is trnsprnt where

the third syllable loses the schwa and the n is a syllabic

consonant Frequentstylistic variants retain the status co de

P for primary pronunciation but always have a pronunci

ation number which is greater than Less frequentstylistic

variants are considered secondary pronunciations suchas

trnsprnt and trnsprnt

Of course there may b e alternative less frequentpronuncia

tions of the primary citation form alternatives whichcantbe

attributed to the word b eing used in connected sp eech Such

dierences sometimes known as sp eakertosp eaker varia

tions remain in all the stylistic variants each dierent ci

tation form has If you ask someone how to pronounce the

word transparentyou might hear the primary form ab ove or

instead a secondary form like trAnsprnt or trn

spErnt the dierence b eing the vowel used in the rst

Status of pronunciation

and second syllables resp ectively When the secondary form

o ccurs in sp eech it might b e pronounced with some stylistic

variation as trAnsprnt or trnspErnt No

matter what the stylistic variations the phoneme which dis

tinguishes one citation form from another remains the same

Secondary forms whether citation forms or stylistic variants

are classied as such b ecause they are thoughttobeused

less commonly than primary forms All secondary forms get

the co de S and all have a pronunciation numb er greater

than Unlike primary forms where the rst form listed

is automatically the citation form secondary forms are not

given in any sort of order So while the rst secondary form

you see listed in the database might b e a citation form it

could just as well b e one of manystylistic variants

Pronunciation Status Pronunciation Example

typ e co de Number passenger

Primary P psIndZr

P psInZr

Secondary S psndZr

S psnZr

S psndZr

S psnZr

Table Pronunciation status co des for English

The flex name and description of this column are as follows

PronStatus Status of pronunciation

PronStatusLemma

PHONETIC TRANSCRIPTIONS

When you b egin selecting phonetic transcriptions from the

celex databases you havetocho ose from a wide arrayof

options First you must decide whether or not you want

to use transcriptions which include syllable b oundaries and

stress markers Then you havetocho ose which of the four

sets of computer phonetic co des b est suits your task and

your p ersonal preferences So b efore dening eachofthe

columns in turn the section b elow describ es the features of

the four available sets of computer phonetic co des In the last subsection after the column denitions there is a table

english linguistic guide

of examples whichgives some transcriptions from eachof

the transcription columns it is esp ecially useful for compar

ing the dierenttyp es of transcription plain syllabied or

stressed and syllabied

COMPUTER PHONETIC CHARACTER SETS

Four dierent sets of phonetic character co des are available

from celex The rst three sets are sampa celex and

cpa and they can b e thought of as computerized versions

of ipa They use standard asci i co desthose whichcanbe

typ ed in and read on almost any terminalto represent cer

tain of the ipa characters As far as p ossible these sets have

b een designed to resemble ipa a lot of the characters you

typ e or read lo ok liketheir ipa counterparts As with ipa

diphthongs and aricates are represented by writing the two

appropriate characters next to each other and long vowels

are indicated by length markers In some cases however

these conventions can lead to ambiguity are the twovowels

shown next to each other really adiphthong or are they

in fact two separate vowels Toovercome such problems

there are columns whichcontain transcriptions with syllable

markers and also columns available whichhave a delimiter

placed after each consonant aricate vowel long vowel or

diphthong So these sets of computer co des for phonetic

transcription can provide a readable approximation of ipa

with extra provision made to overcome the p ossibilityof

ambiguity

The tables over the next two pages list the basic set of

segments for English Each line gives an ipa character along

side a word which exemplies the sound and the equivalent

characters in the four computerusable sets available with

celex

The rst of the three ipalike sets is the sampa set It

was develop ed in connection with a Europ ean Community

research program and it has b een presented in the Jour

nal of the International Phonetic Asso ciation

pp as a widelyagreed computerreadable phonetic

character set suitable for use with Danish Dutch English

French German and Italian For technical reasons the ver

sion of sampa implementedby celex has to include one

change the character asci i co de representing the

Computer phonetic character sets

ipa example sampa celex cpa disc

p pat p p p p

b bad b b b b

t tack t t t t

d dad d d d d

k cad k k k k

game g g g g

bang N N N N

m mad m m m m

n nat n n n n

l lad l l l l

r rat r r r r

f fat f f f f

v vat v v v v

S thin T T T T

then D D D D

s sap s s s s

z zap z z z z

M sheep S S S S

measure Z Z Z Z

j yank j j j j

x loch x x x x

h had h h h h

w why w w w w

Q cheap tS tS T J

jeep dZ dZ J

bacon N N N C

j

m idealism m m m F

j

n burden n n n H

j

l dangle l l l P

j

father r r r R

p ossible linking r

Table Computer phonetic co des for English consonants

halfop en front rounded vowel sound has b een implemented

as asci i co de The second is a set originally designed

for use within celex The third is cpatheComputer

Phonetic Alphab etorEsprit whichwas develop ed in

the Ruhr Universitat Bo chum Germany

The fourth set is the disc set so called b ecause it is a

computer phonetic alphab et made up of distinct single char

acters It is fundamentally dierent from the other three in

english linguistic guide

ipa example sampa celex cpa disc

pit I I I I

pet E E E E

pat

putt V V V

pot Q O O Q

V put U U U U

another

iq bean i i i i

q barn A A A

q born O O O

uq boon u u u u

q burn

e bay eI eI e

a buy aI aI a

boy OI OI o

V no U U O

aV brow aU aU A

peer I I I

pair E E E

V poor U U U

timbre c

q detente A A A q

q lingerie

q bouillon O O O

Table Computer phonetic co des for English vowels and diphthongs

that it assigns one asci i co de to each distinct phonological

segment in the sound systems of Dutch English and Ger

man Here segment means a consonant an aricate a short

vowel a long vowel or a diphthong There are two main

advantages to this set First it provides one character for

one segment in contrast to the other three sets which use

extra characters for long vowels aricates and diphthongs

Second there is no p ossibilityofambiguous transcriptions A

diphthong is always shown as a diphthong and two separate

vowels in proximitytoeach other say on either side of a

syllable b oundary can thus no longer b e confused with a

real diphthong an aricate is always shown as such and not

as two consonants For b oth these reasons those interested

in pro cessing phonetic transcriptionsas opp osed to reading

transcriptions in a character set that resembles the familiar

Computer phonetic character sets

ipamaywell cho ose transcriptions in this character set

Its most basic co des corresp ond to sampa all the sam

pa co des which represent short vowels and consonants are

included in this set The remaining long vowels diphthongs

and aricates have b een assigned co des not already in use

for other purp oses The resulting character set thus do es not

lo ok as elegantand ipalike as the other three sets However

if you are mainly interested in the computer pro cessing of

transcriptions such sthetic considerations might not b e so

imp ortant

Clearlyyou have a wide choice of transcriptions available

to you The typ e you cho ose will dep end on the nature

of the task you have in mind For ipalike readabilityand

nonambiguous transcriptions use the sampa celex or

cpa sets For computer pro cessing tasks which need one

charactertoonesegmentcorresp ondence use the disc set

In App endix I there is a table which sets out disc and how

it relates to Dutch English and German One nal p oint

worth noting is that if instead of the standard sets of co des

oered here you want to use a set of your own making

you can implementitby means of the pattern transducer

available in the flex window MODIFY CONVERSION and the

disc character set is probably the easiest set to convert

PLAIN TRANSCRIPTIONS

The rst set of columns oers plain transcriptions that

is transcriptions which do not haveany syllable markers or

stress markers written in each of the four co ding systems

already describ ed However three of these columns have one

sp ecial feature each phonetic segment ends with a delimiter

Here a segment means a vowel a consonant a long vowel

a diphthong or an aricate Using a delimiter avoids any

p ossibilityofambiguitybetween the two parts of a diphthong

or an aricate These delimiter transcriptions are available in

the sampa celexandcpa characters sets Delimiters are

not given with disc transcriptions since the unique single

character nature of that set obviates the need to delimit each

segmentinthiswayAlsoavailable with the plain transcrip

tions is a column which indicates howmany phonemes each

transcription contains

The rst plain transcription column uses the sampa charac

ter set and full stops as delimiters The flex name and

english linguistic guide

description of this column are as follows

PhonSAM Unsyllabified SAMPA character set

PhonSAMLemma

The second column uses the celex character set and full

stops as delimiters The flex name and description of

this column is as follows

PhonCLX Unsyllabified CELEX character set

PhonCLXLemma

The third column uses the cpa character set and full stops

as delimiters Normally cpa uses full stops as syllable

markers but here of course no syllable markers are used

The flex name and description of this column is as follows

PhonCPA Unsyllabified CPA character set

PhonCPALemma

The last plain transcription column uses the disc set No

delimiters syllable markers or stress markers are included

The flex name and description of this column are as follows

Unsyllabified DISC character set PhonDISC

PhonDISCLemma

A countwhich tells you howmany phonemes each headword

you select contains is available Since certain phonemes long

vowels and diphthongs are sometimes given bytwocharac

ters this count is more sophisticated than merely the length

of the string Here are some examples the lemma farmhouse

has six phonemes in its transcription famhaUsand

ammeter also has six mItr

PhonCnt Numbern of phonemes

PhonCntLemma

Syllabied transcriptio ns

SYLLABIFIED TRANSCRIPTIONS

The next set of transcriptions use the same basic transcrip

tions as the plain set but this time they are given without

any phoneme delimiters Instead the syllables whichmake

up eachword are shown in one of twoways The rst metho d

is to use a hyphen or in the case of cpa a full stop to

mark every syllable b oundary within eachword The second

metho d available with the celex character set is to enclose

each syllable within square brackets

The rst syllabied transcription column uses the sampa

character set It uses a hyphen as the syllable marker and

its flex name and description are as follows

PhonSylSAM Syllabified SAMPA character set

PhonSylSAMLemma

The next two syllabied transcription columns use the celex

character set and the rst of these uses hyphens to mark

every syllable b oundary in eachword The flex name and

description of this column are as follows

Syllabified CELEX character set PhonSylCLX

PhonSylCLXLemma

The other celex syllabied column uses the brackets no

tation as describ ed ab ove and its flex name and description

are as follows

PhonSylBCLX Syllabified CELEX character set brackets

PhonSylBCLXLemma

The fourth syllabied transcription column uses the cpa

character set and every syllable b oundary within eachword

is marked by a full stop The flex name and description of

this column are as follows

PhonSylCPA Syllabified CPA character set

PhonSylCPALemma

english linguistic guide

The fth syllabied transcription column uses the character

set called discandevery syllable b oundary within eachword

is marked byahyphen The flex name and description of

this column are as follows

PhonSylDISC Syllabified DISC character set

PhonSylDISCLemma

Thelastcolumninthissetgives for each lemma or wordform

a count whichtellsyou the numb er of phonetic syllables in

the word For example famhaUs contains two sylla

bles and mItr contains three

SylCnt Number of phonetic syllables

SylCntLemma

STRESSED AND SYLLABIFIED TRANSCRIPTIONS

The third set of columns gives transcriptions which are syl

labied and also have primary and secondary stress markers

There are four such columns eachcontaining transcriptions

in one of the four computer usable phonetic character sets

describ ed ab ove

The rst column uses the sampa character set and as well

as using hyphens to mark syllable b oundaries these tran

scriptions show p oints of primary stress by means of the

double quote character and p oints of secondary stress

by means of the p ercent character These characters

are placed immediately b efore a stressed syllable The flex

name and description of this column are as follows

PhonStrsSAM Syllabified with stress marker SAMPA

PhonStrsSAMLemma character set

The second column uses the celex character set and as

well as using hyphens to mark syllable b oundaries these

transcriptions showthepoints of primary stress with an

inverted comma and the p oints of secondary stress with

a double quote

PhonStrsCLX Syllabified with stress marker CELEX

PhonStrsCLXLemma character set

Stressed and syllabied transcriptio ns

The third column uses the cpa character set including full

stops to mark syllable b oundaries these transcriptions show

p oints of primary stress with an inverted comma and

the p oints of secondary stress with a double quote

PhonStrsCPA Syllabified with stress marker CPA

PhonStrsCPALemma character set

The fourth column uses the disc character set along with

hyphens to mark syllable b oundaries these transcriptions

show p oints of primary stress with an inverted comma

and p oints of secondary stress with a double quote

PhonStrsDISC Syllabified with stress marker DISC

PhonStrsDISCLemma character set

A stress pattern is available for each lemma or wordform as

well as a countofthenumb er of syllables and phonemes each

contains

A stress pattern is a string whichshows how each syllable

is stressed in sp eech Each syllable is represented by one

numeric character either or indicates that the

syllable receives secondary stress indicates that it receives

primary stress and that it do es not receive primary or

secondary stress The examples b elow contrast the syllabi

ed phonetic transcription which includes a primary stress

marker with the stress pattern describ ed here

Example Transcription Stress pattern

biographic baIUgrfIk

go ogly guglI

Table Example stress patterns

StrsPat Stress pattern

StrsPatLemma

english linguistic guide

EXAMPLE TRANSCRIPTIONS

Column Examples

excruciating ly o ceanic

Plain transcriptions

PhonSAM IkskruS IeI tI Nl I USInIk

PhonCLX IkskruS IeI tI Nl I USInIk

PhonCPA IkskruS Ie tI Nl I OSInIk

PhonDISC IkskruSItINlI SInIk

Syllabied transcriptio ns

PhonSylSAM IkskruSIeI tIN lI USInIk

PhonSylCLX IkskruSIeI tIN lI USInIk

PhonSylBCLX IkskruSI eI tIN lI USInIk

PhonSylCPA IkskruSIe tIN lI OSInIk

PhonSylDISC IkskruSIt INlI SInIk

Syllabied and stressed transcription s

PhonStrsSAM IkskruSIe ItIN lI USInIk

PhonStrsCLX IkskruSIe ItIN lI USInIk

PhonStrsCPA IkskruSIe tIN lI OSInIk

PhonStrsDISC IkskruSI tINl I SInIk

Table Example English Phonetic Transcriptio ns

PHONETIC PATTERNS

Phonetic patterns here means cv patterns the consonant

and vowel patterns for the phonetic transcription as op

p osed to the orthographic transcriptions of any headword

or wordform you select Instead of the basic cv pattern

whichuseshyphens to mark phonetic syllable b oundaries

within words you maywant to use the alternative notation

which delimits syllables by means of square brackets The

phonetic cv patterns used here representeach short vowel

as Veach long vowel and diphthong as VVeach consonant

and aricate as C and each syllabic consonantas S

This table illustrates the two dierent formats you can cho ose

for your cv patterns

Phonetic patterns

Example Transcriptio n CV pattern CV pattern

with brackets

farmhouse famhaUs CVVCCVVC CVVCCVVC

ammeter mItr VCVCVC VCVCVC

Table Example CV patterns

The basic phonetic cv patterns include hyphens as syllable

markers The flex name and description of this column are

as follows

PhonCV Phonetic CV pattern

PhonCVLemma

Alternatively you can cho ose phonetic cv patterns of head

words which use square brackets to delimit the syllables

This column has the following flex name and description

Phonetic CV pattern with brackets PhonCVBr PhonCVBrLemma

english linguistic guide

ENGLISH MORPHOLOGY

Information on English Morphology is available with lemma

lexicons and wordform lexicons If you are interested in

inectional morphologythenyou should use a wordforms

lexicon and if you are interested in derivational and com

p ositional morphologyyou should use a lemma lexicon

MORPHOLOGY OF ENGLISH LEMMAS

The morphological analyses given for lemmas in the celex

databases always use the headword form of the lemma b e

cause this form unlikeDutch is usually the shortest in any

inectional paradigm without any visible inectional end

ings However when discussing English morphology stem is

the normal term used to describ e this form and so in this

section stem is used instead of headwordjusttotinwith

common practice

Before nding out details ab out each of the columns avail

able you should lo ok at the sections b elowwhich try to give

some explanation of the metho ds used to obtain the analyses

given in the database You will then know what celex

means by terms suchasimmediate segmentation hierar

chical segmentation comp ound derivation and derivational

comp ound After all that youll understand more clearly

what eachofthevarious columns has to oer

HOW TO SEGMENT A STEM

The rst and most fundamental typ e of segmentation is im

mediate segmentation This simply involves splitting a stem

into its largest constituent parts If you continue to carry

out immediate segmentation until there is nothing left to

segment you arrive at the stems complete segmentation

Dep ending on your requirements you can lo ok at a complete

segmentation in two forms The rst is the at form which

shows every morpheme that makes up the stem The second

is the hierarchical form which as well as p ointing out the

individual morphemes in a stem also shows all the analyses

How to segmentastem

whichhave to b e made to identify those morphemes The

at segmentation gives the conclusion reached while the

hierarchical segmentation shows the working

To illustrate the three typ es of segmentation takeasan

example the word interdenominational

The rst typ e of analysis Immediate segmentation gives the

ax inter plus the stem denominational

interdenominati ona l

inter denomination al

The second typ e of analysis complete segmentation at

shows you what you get if you keep applying immediate

segmentation namely the constituent morphemes of inter

denominational the ax inter plus the ax de plus the

stem nominate plus the ax ion plus the ax al

interdenominati ona l

inter de nominate ion al

The third typ e complete segmentation hierarchical shows

you the full analysis of the word including each individual

immediate segmentation carried out It gives you enough

information to pro duce a hierarchical tree diagram like this

one

interdenominat iona l

denomination al

denomination

denominate

inter de nominate ion al

english linguistic guide

For most stems in the database representations of eachof

these three typ es of segmentation are available Sometimes

there is more than one representation b ecause certain stems

can have more than one immediate segmentation To explain

this fully the next section describ es the basic analyses that

result from immediate segmentation

TYPES OF ANALYSES

When you attempt to split a stem into its biggest comp onent

parts the result is always some combination of stems or ec

tions and axes A ectionsuchasthefreezing in freezing

p ointis treated the same as a stem so that whenever an

analysis involves a stem you know that the stem could also

b e a ection The most straightforward analysis of all is

a stem which consists of only one free morpheme it is

monomorphemic and clearly cant b e split up Every other

stem however consists of one smaller stem or ax plus at

least one ax or one other stem and can b e termed either

a Derivationa Comp oundaDerivational Comp oundora

Neoclassical Comp ound It is imp ortant to understand the

dierences b etween these four terms since they are at the

heart of the morphological information celex provides So

in the subsections b elow each is dened in terms of stems

and axes Examples are given and simple tree diagrams

illustrate the appropriate immediate analyses

THE DERIVATION

A derivation involves axation whereby axes can b e

added to an existing stem or ection to form a new stem

The immediate analysis always takes one of four p ossible

forms

i a binary split into a stem or ection plus an ax the

word careful for example care ful

stem

stem affix

The Derivation

ii a binary split into an ax plus a stem or ection For ex

ample the word barometer is analysed as baro meter

stem

affix stem

iii a triform split into an ax a stem or ection and an

ax the word extracurricular for example extra curricu

lumar

stem

affix stem affix

iv a triform split into a stem or ection an ax and another

ax Suchwords can b e derivations of inected forms

the word falteringly for example is analysed as falter

ing ly which is a stem plus an inectional ax plus

a derivational ax Alternatively they can b e lexicalised

forms of inected derivations like countried analysed as

country ify ed which is a stem plus a derivational

ax plus an inectional ax This sort of analysis is only

appropriate when the stem and the ax which immediately

follows it dont together form a lemma b ecause otherwise

as with the word inationarythe immediate analysis would

b e liketyp e i ab ove a stem plus an ax ination ary

stem

stem affix affix

english linguistic guide

THE COMPOUND

A compound is the joining of two stems or ections into

one new stem The immediate analysis always takes one of

two forms

i a binary split into two stems the word nameplate for

example name plate

stem

stem stem

ii a triform split intoastemorectionanaxsimplya

link morpheme and a stem or ection the word bandsman

for example band s man

stem

stem affix stem

Words which consist of more than two stems arent ana

lysed as comp ounds since they normally have the structure

of a phrase or sentence So headwords like nevertheless

Australian Rules fo otball and b eallandendall dont get

a morphological analysis

THE DERIVATIONAL COMPOUND

A derivational compound is a comp ound which can only

b e formed in combination with a derivational ax as op

p osed to a simple link morpheme The immediate analysis

normally takes the form of a triform split into a stem or

ection another stem or ection and an ax the word

icebreaker for example ice break er

stem

stem stem affix

The Derivational Comp ound

A couple of words can b e analysed as a quaternary split into a

stem an ax a stem and an ax the word brinksmanship

for example is analysed as brinksmanship and

whipp ersnapp er is analysed as whipersnaper

However this is a very rare form of analysis

stem

stem affix stem affix

THE NEOCLASSICAL COMPOUND

A neoclassical compound is a word which app ears to b e

made up of two axes neither of which can o ccur as a word

in its own right like aero drome aero drome orneurology

neuro ology The axes which combine to form this typ e

of comp ound are generally known as combining forms

stem

affix affix

THE NOUNVERBAFFIX COMPOUND

Problems sometimes arise with the analysis of words which

lo ok likederivational comp ounds The general denition of a

derivational comp ound is normally sucient but when the

secondstemisaverbal form things b ecome more compli

cated A stem which comprises a noun plus a verb plus an af

x can normally b e considered a derivational comp ound but

some p eople maywant to treat it as an ordinary comp ound

or derivation The distinction is imp ortant since it can aect

not only the app earance of a single immediate segmentation

branch but also the app earance of a complete hierarchical

tree The stem copy editor is such a problem comp ound If

you consider it to b e an ordinary comp ound the stem copy

english linguistic guide

plus the stem editor its complete hierarchical tree lo oks like

this

copy editor

copy editor

edit or

If you consider it to b e an ordinary derivation the stem copy

edit plus the ax or its complete hierarchical tree lo oks

like this

copy editor

copyedit or

copy edit

But if you consider it to b e a derivational comp ound the rst

immediate segmentation gives you the stem copy plus the

stem edit plus the ax orwhichgives the full hierarchical

tree a dierent app earance

copy editor

copy edit or

HOWTOASSIGNANANALYSIS

When youre faced with a headword that needs to b e ana

lysed howdoyou work out the correct analysis How did the

p eople at celex who carried out the morphological analysis

by hand arrive at the answers contained in the database

In particular in the case of nounplusverbplusax words

How to assign an analysis

how did they decide which of the analysis typ es discussed in

the previous section were appropriate

To illustrate the principles used in analysing the information

there are two diagrams given as Tables and b elow The

rst illustrates the general strategy adopted for each head

word and the second deals with the sp ecial problems that

arise with nounplusverbplusax words In b oth diagrams

abbreviations are used s means stemanda means ax

making it easy to refer back to the sections ab ove which

dene derivations comp ounds and derivational comp ounds

in terms of stems and axes When an analysis is acceptable

it means that the comp onent parts identied are current

stems or axes and that the word can b e dened as a deri

vation a comp ound or a derivational comp ound according

to the denitions given in sections ab ove An

acceptable stem is one which app ears in the Collins English

Dictionary without b eing marked as obsolete or archaic

Following the rst diagram analysis starts with an attempt

to see if the word under scrutiny is just the same as an already

existing word with a dierentword class The word railroad

for example can b e used as a verb and it is said to come from

the corresp onding noun railroad This phenomenon is called

conversion or zero derivation since there is no dierence in

the form of the twowords even though they have a dierent

word class Conversion is explained in full under section

Status and Language co des If conversion has o ccurred the

analysis need go no further in MorphStatus the word gets

the co de Z and NVAComp and its sub ordinate columns

Der Comp and DerComp are all set to N

If the word is not a conversion then the next step is to check

whether it ts with the denition of a derivation given in

section ab ove For example the word calculator is

analysed as the stem calculate with the sux or encircle is

analysed as the prex en with the stem circle and unap

pable as the prex un plus the stem ap plus the sux able

In all three cases the word is classied as a derivationin

MorphStatus the word gets the co de C to indicate that it

is complexand NVAComp and its sub ordinate columns

Der Comp and DerComp are all set to N

If the word turns out not to b e a derivation then the next

stage is to see if it ts with the denition of a comp ound given

english linguistic guide

Is the lemma derived from another

lemma identical in form but

dierentinword class

yes no

Zero Derivation Is the analysis sa or as or asa

acceptable

yes no

Derivation Is the analysis ss or sas acceptable

yes no

Comp ound Is the analysis ssa or sasa

acceptable

yes no

Derivational Monomorphemic

Comp ound

Table How to carry out morphological analysis

in section ab ove For example the noun keyb oard is

analysed as the stem key plus the stem b oardandgrounds

man as the stem ground plus the inx s plus the stem man

In b oth cases the word is classied as a comp ound under

the MorphStatus column it gets the co de C to indicate

that it is complexand NVAComp and its sub ordinate

columns Der Comp and DerComp are all set to N

If the word still hasnt b een classied then the last stage

is to see whether the word is a derivational comp oundas

dened in section ab ove For example the adjective

barefaced is analysed as the stem bare plus the stem face

plus the ax ed The word is therefore classied as a

derivational comp ound under MorphStatus it gets the

co de C to indicate that it is complexandNVAComp

and its sub ordinate columns Der Comp and DerComp

How to assign an analysis

are all set to N

It is p ossible that the word might not t into anyofthe

ab ove categories this makes the word monomorphemicand

the analysis is simply the word itself chair for example

or llama Under MorphStatus the co de M is given and

NVAComp and its sub ordinate columns Der Comp

and DerComp are all set to N In other cases where no

analysis can b e carried out the co de under MorphStatus

indicates why You can read ab out these co des in section

Status and language co des

THE NOUNVERBAFFIX COMPOUND

The general scheme explained ab oveisenoughtoarriveat

an analysis in most cases However diculties in apply

ing the system arise when you start considering socalled

nounverbax comp ounds those words whichcontain a

verbal element which arent conversions and which could

b e analysed as a nominal stem plus a verbal stem plus an

ax Examples of suchwords are sto ckholder and copy

editorThistyp e of comp ound is characterised bya Y in

the NVAComp column As the diagram b elowshows

just b ecause they could b e analysed in suchaway it do esnt

mean they necessarily should be Sto ckholder is b oth a

comp ound and a derivational comp ound and copyeditor is

a derivation a comp ound and a derivational comp ound The

approach outlined b elow is designed to keep as many mor

phologists as p ossible happy with the information available

in the database its p ossible to cho ose for yourself whether

to restrict your lexicon to just one typ e of analysis or to

p ermit them all according to your own requirements

The rst step is to see whether the NVAComp word

you are dealing with can b e classied as a derivation in

accordance with the denition in section ab ove Take

the word diveb omber as an example it can b e analysed as

the stem diveb omb plus the ax er The verb diveb omb

is accepted as legitimate b ecause it o ccurs as suchinthe

Collins English Dictionary ced So this word do es meet

the denition of a derivation and thus gets the co de Y in

the Der column Another example is the word bricklayer

since a verb bricklay do esnt exist according to the cedit

cant b e analysed as a stem plus an ax and so it gets the

co de N in the Der column

english linguistic guide

Is the analysis sa or as or asa

acceptable

yes no

Derivation Is the analysis ss or sas acceptable

yes no

Comp ound Is the analysis ssa or sasa

acceptable

yes no

Derivational

Comp ound

Table Dealing with nounverba x comp ound analyses

The next stage is to see whether the NVAComp word in

question can b e classied as a comp ound as dened in sec

tion ab ove Again the word diveb omber meets the

denition it can b e analysed as the stem dive plus the stem

bomber and it is a particular sort of b omb er It can therefore

b e called a comp ound and it gets the co de Y in the Comp

column Applying the same rules to bricklayer pro duces the

opp osite result a bricklayer isnt really a particular sort

of layer since according to the ced layer do esnt mean

someone who lays things So bricklayer gets the co de N in

the Comp column to show that it isnt a comp ound

The last stage is to decide whether the word is a derivational

comp ound as dened in section ab ove This time

diveb omber do es not qualifyEven though it can havethe

structure stem dive plus stem bomb plus ax er the

The NounVerbA x Comp ound

noun dive cannot b e the ob ject of the verb bomb you cant

talk ab out b ombing a dive Since dive isnt any sort of

obligatory complement to the verb bomb diveb omber is

not a derivational comp ound and therefore gets the co de N

in the DerComp column Applying the rules to bricklayer

again pro duces the opp osite result It can have the structure

stem brick plus stem lay plus ax er and brick is the

ob ject of the verb lay it is p ossible to talk ab out laying

bricks So since this time bricks is some sort of complement

to the verb lay bricklayer gets the co de Y in DerComp

column to show that it is a derivational comp ound

To illustrate this pro cess further more examples are given in

the table b elow

Word Classicati ons

NVAComp Der Comp DerComp

typ esetter Y Y N Y

diveb omber Y Y Y N

copyeditor Y Y Y Y

sto ckholder Y N Y Y

churchgo er Y N N Y

cub rep orter Y N Y N

Table Example nounverba x comp ound analyses

It shows six examples and the co des each one gets in NVA

Comp Der CompandDerComp as a quickwayof

showing howthewords are classied in the NVAComp

analysis scheme In the database however each separate

analysis gets a separate row and if you lo oked up analyses

for these six words you would get something like this

english linguistic guide

Stem MorphNum NVAComp Der Comp DerComp Def Imm

typesetter Y Y N N Y typeseter

typesetter Y N N Y Y typeseter

divebomber Y Y N N Y divebomber

divebomber Y N Y N Y divebomber

copyeditor Y Y N N Y copyeditor

copyeditor Y N Y N Y copyeditor

copyeditor Y N N Y Y copyeditor

stockholder Y N Y N Y stockholder

stockholder Y N N Y Y stockholder

churchgoer Y N N Y Y churchgoer

cub reporter Y N Y N Y cubreporter

If youve followed in full the explanation of how celex car

ried out its morphological analysis of English words most

of this example lexicon should b e clear The columns it

contains along with other columns are describ ed and dened

in the sections that follow Using the columns available you

can control the numb er of analyses you see for each stem as

well as the typ e of analyses by means of restrictions on the

numb er and status columns which are dened b elow You

can decide for yourself whether your lexicon should contain

just one default analysis p er stem or whether it should

contain more than one analysis p er stem In cases where

a stem can b e analysed as a derivation a comp ound or a

derivational comp ound you can cho ose to include whichever

typ e you prefer leaving out the other typ e In short you

have the freedom to build lexicons whichcontain morpho

logical information in the form you most prefer

STATUS AND LANGUAGE CODES

The rst ADD COLUMNS menuyou see after you select the Morphology option is this one

Status and language co des

ADD COLUMNS

Status

Language information

Derivationalcom posit iona l information

TOP MENU

PREVIOUS MENU

Before dealing with the various derivationalcomp ositional

information columns which form the bulk of the available

morphological information the rst two columns are dealt

with here

The rst column simply tells you by means of a single co de

whether each stem is morphologically simple morphologi

cally complex or a conversion or whyitisasyet unanalysed

The table b elow shows the co des that are used and it is

followed by a description of each of the eight co des Just

b efore the description concludes with the column denition

there is a diagram which illustrates the strategies celex used

to determine a status co de for each stem

Status Co de Example

Morphological analysis available

Morphologically complex C sandbank

Monomorphemic M camel

Conversion Zero Derivation Z abandon

Contracted form F Ive

Morphological analysis unavailable

Morphology irrelevant I meow

Morphology obscure O dedicate

Morphology may include a ro ot R imprimatur

Morphology undetermined U hinterland

Table Derivational morphology status co des

If a stem contains at least one stem plus at least one other

stem or ax then it is said to b e morphologically complex

Details of how the stem can b e analysed are given in the

derivationalcomp ositional segmentation columns describ ed

english linguistic guide

in the section b elow Thus if a stem has the morphological

status co de C for complex you know that information

ab out its derivational andor comp ositional morphology are

available in the database

If a stem is monomorphemic then it contains only one mor

pheme and no further analysis is required The morphologi

cal status co de M means monomorphemic and you know

that a simple onestem analysis is given as the derivational

andor comp ositional morphology for each stem with this

co de

If a stem app ears to b e derived from another stem whichis

identical in form but dierentinword class it gets the co de Z

for zero derivation or conversion The noun delinquentfor

example can b e said to derive from the adjective delinquent

Normally derivations from one word class to another are

clearly marked by means of an ax sheepish is an adjective

derived from the noun sheep for example But conversions

on the other hand are not so marked itsasifanax

containing nothing had b een added to the original stem

Naturally enough when conversion o ccurs its not immedi

ately obvious which stem is the original and whichisthe

derivative In analysing these words celex adopted a strat

egy for determining the direction of the conversion that is

if a verb has b een converted into a noun the derivation is in

the direction of the noun Table indicates the normal

direction of conversion Conversion in the opp osite direction

is also p ossible provided that it is sp ecied in the Shorter

Oxford English Dictionary soed

Default direction Example

verbnoun paint

adjectivenoun parallel

adjectiveadverb pretty

adjectiveverb pale

prepositionadverb past

Table Direction of conversions

The status co de F indicates that the analysis given is in fact

acontraction The single contraction d can represent had

would or did and the complex contraction hes represents he

Status and language co des

is or he hasEachcontracted form gets its own rowinthe

database and the status co de F

In the case of monomorphemic stems complex stems conver

sion stems and contractions of stems morphological analyses

are provided in the various segmentation columns However

there remains a large numb er of stems whichhave no analy

sis and in such cases co des indicating the reasons for the

lack of analysis are given in this column and these co des

and reasons are explained b elow

First of all sometimes even attempting morphological analy

sis is not appropriate for a particular stem Usually this is

true when the stem is an exclamation or an interjection of

some sort gosh prithee or meow for example or when

it is a prop er noun Spooner and Germany arent ana

lysed In addition those few words which seem to have

taken on the structure of a short sentence or at least consist

of three or more stems like nowadays or who dunit dont

get an analysis So whenever a stem has the co de I for

irrelevant you know that a morphological analysis isnt

considered necessary and that its entries in the segmentation

columns describ ed b elow are therefore empty

Some stems are recognizable recentloanwords whichhave

achieved some sort of currency in English words like virtu

oso or pretzel or mazurka Since providing analyses for such

stems would in many cases mean delving into the morphol

ogy of languages not covered by celex they simply receive

the co de U for undetermined The languages loanwords

originate from are shown in the next column Lang

On other o ccasions an analysis seems p ossible but cannot

b e fully explained The stem tabby for instance app ears to

consist of the pro ductivesuxy plus what might b e another

stem tabHowever tab b ears no immediate relation to the

adjective tabbysothattabby gets the co de O to indicate

that the morphological analysis is obscure

In most cases morphological analysis is carried out on a

synchronic basis the stems or axes whichmakeupa

word must o ccur in mo dern current English regardless of

the historical origins they mighthave On many o ccasions

however an etymological ro ot could explain the morphology

of a stem whichwould otherwise b e unanalysable The stem

english linguistic guide

patrimony for example app ears to b e made up of a Latin

prex patri andwhatmay b e a Latin sux monyStems

like this which could b e analysed on the basis of the histor

ical ro ot of its constituent parts are given the co de R for

ro ot

If you want to understand this co ding system more fully then

you can examine Table which is a diagram that sets out

ascheme for arriving at appropriate morphological status

co des Starting with a stem whose morphology is relevant

that is it do esnt b elong with those stems that have the co de

I you can work out the co de it should havebyfollowing

the diagram through This strategy is the one actually used

by celex to determine the correct co des

This column can b e used to eliminate from your lexicon stems

for which there are no morphological analyses allowing you

to concentrate on those which do Simply add a restriction

which states that you only want stems which are morpholog

ically complex MorphStatus C

The column whichcontains these morphological status co des

has the following flex name and description

MorphStatus Morphological status

MorphStatusLemma

The second column contains co des that identify the par

ticular geographical origins of lemmas including the foreign

languages from which some of the lemmas havebeenbor

rowed The headword conquistador thus gets the co de S to

indicate that it is of Spanish origin and the verb wae gets

the co de B b ecause its reckoned to b e a p eculiarly British

usage Whenever a lemma do esnt have a co de it isnt a

recent b orrowing from another language and it do esnt have

asso ciations with one regional variety of English Table

below sets out all the co des used and their meanings

Status and language co des

Is the Lemma derived from another

lemma identical in form but

dierentinword class

yes no

Z Do es the Lemma contain at least

Zero derivation one mo dern pro ductive lemma

or ax plus one other element

yes no

Do es the other Is the Lemma a

elementcontain at recentloanword

least one pro ductive

lemma or ax

yes no yes no

C O U Do es the Lemma

Complex Obscure Undetermined contain at least

one etymological ly

recognisable ro ot

yes no

R M

Ro ot Monomorphemic

Table How to assign MorphStatus co des

english linguistic guide

Language Co de Meaning Example

A American English billfold

F French patisserie

B British English divvy

D German sauerkraut

G Greek eureka

I Italian cicerone

L Latin emeritus

S Spanish siesta

Table Language co des for English headwords

The flex name and description of the column which contains

these co des is as follows

Lang Language information

LangLemma

DERIVATIONALCOMPOSITIONAL INFORMATION

ADD COLUMNS

Number of morphological analyses

Analysis number N

Status of morphological analysis

Segmentations

Other

TOP MENU

PREVIOUS MENU

These options giveyou information ab out the derivational

and comp ositional morphology of stems including howmany

analyses are available for each stem a unique numb er for each

analysis an indication of the wayinwhicheach analysis has

b een made and a marker for the default analyses for each

stem

The rst option is a column which simply indicates howmany

analyses have b een made for each stem For example back

re has one analysis ashbulb has two and treasurer three

The numb er of analyses for each stem also equals the number

Derivationalcom p os ition al information

of rows that stem can have with distinct analyses since each

morphological analysis is assigned to its own individual row

You can use this column to construct restrictions for your

lexicon A simple example would b e one that includes in your

lexicon only those stems whichhave more than one analysis

This would take the form MorphCnt The flex name

and description of this column are as follows

MorphCnt Number of morphological analyses

MorphCntLemma

The second option is a column which identies each analysis

of a particular stem Each dierent morphological analysis

of a stem is assigned to a dierentrow and this column gives

the number of the row Thus the adjective lemma ashbulb

has tworows one has the MorphNum the other has the

MorphNum or a stem description of this column are as

follows

Morphological analysis ID MorphNum

MorphNumLemma

ANALYSIS TYPE CODES

Under the status of morphological analysis option there

are veyesnotyp e columns which when you use them

to construct restrictions can help you extract the analyses

you want from the many stem segmentations available

Each distinct morphological analysis of eachstemhasa num

b er and is given in several dierent forms on its own rowin

the database These columns give simple information ab out

each analysis and are particularly useful whenever a stem is

a nounverbax comp ound A nounverbax comp ound

as discussed in section can correctly b e analysed

as a derivation a comp ound or a derivational comp ound

The ve columns in question are called NVAComp Der

Comp DerCompand Def

Whenever NVAComp contains a Yyou know that yes

this rowcontains a stem which is considered a nounverb

ax comp ound and which therefore might b e analysed in

three dierentways And naturally whenever it contains

english linguistic guide

an Nyou know that the rowcontains a stem whichis not

considered a nounverbax comp ound The flex name and

description of this column are as follows

NVAComp Nounverbaffix compound

NVACompLemma

Whenever Der contains a Yyou knowthatYes this row

contains a nounverbax comp ound which is analysed as a

derivation And whenever it contains an Nyou know that

the rowcontains a nounverbax comp ound whichisnot

analysed as a derivation or a stem which is not of the noun

verbax comp ound typ e The flex name and description

of this column are as follows

Der Derivation analysis

DerLemma

Whenever Comp contains a Yyou knowthatyes this

row contains a nounverbax comp ound which is analysed

as a comp ound And again N means that the rowcon

tains a nounverbax comp ound which isnt analysed as

a comp ound or a stem which is not of the nounverbax

comp ound typ e The flex name and description of this

column are as follows

Comp Compound analysis

CompLemma

Likewise whenever DerComp contains a Yyou know that

yes this rowcontains a nounverbax comp ound which

is analysed as a derivational comp ound And naturally N

means that the nounverbax comp ound isnt analysed as

a derivational comp ound or that it is a stem which is not

of the nounverbax comp ound typ e The flex name and

description of this column are as follows

DerComp Derivational compound analysis

DerCompLemma

If a stem has more than one analysis its sometimes helpful

to b e able to identify one which is the b est or most useful

or at least to discard unwanted alternatives Whenever Def

Analysis typ e co des

contains a Yyou knowthatyes this row contains a default

analysis and when it contains an Nyou know that the row

contains another nondefault analysis

Since there are three typ es of analyses which can b e assigned

to a complex stem there might also b e up to three default

analyses for one word a default derivation analysis a default

comp ound analysis and a default derivational comp ound

analysis Of course not manywords are eligible for three

default analyses While morphological analysis was b eing

carried out rules were formulated to determine which analy

ses should take precedence over the others and these rules

are explained in Table b elow

The lefthand column gives the problem which those doing

the analysis came up against the central column shows which

of the two p ossible analyses should b e the default or take

precedence and the righthand column illustrates the prin

ciple with an example The rst part of the table shows the

preferential order for derivations and the second part shows

the preferential order for comp ounds A part for derivational

comp ounds isnt necessary since they are only analysed in

one way

If despite the range of analyses available you only want just

one default analysis then you can get it by making a restric

tion on MorphNum MorphNum The rst analysis for

a lemma is always a default analysis Analyses which are

derivations take precedence over comp ounds and likewise

comp ounds take precedence over derivational comp ounds

Using this column in conjunction with the three preceeding

columns you can construct restrictions which select or omit

the analyses you sp ecifyTheflex name and description of

this column are as follows

Def Default analysis

DefLemma

To illustrate howyou can use these columns imagine that

you havechosen Imm and ImmClass as the form of mor

phological analysis you wanttosee Imm shows the analy

sis and ImmClass shows the word class of the analysed

parts these columns and the other columns containing the

same analyses in dierent forms are describ ed in the sections

english linguistic guide

Option Solution Example

Preferential order for the analysis of derivations

stem ax stem ax takes precedence disavowal is analysed rst as

or over ax stem disavow althenasdis

ax stem avowal

verb ending in ate ete ote or The verb with the higher annunciation is analysed rst as

ute the ax ion frequency in the cobuild typ e the verb announce plus the ax

or list takes precedence iation and then as the verb

verb not ending in ate ete ote annunciate plus the ax ion

or ute an ax like tion

problematical ly is analysed rst The adjective with the higher adjectivesuxly

as problematic allyand frequency in the cobuild typ e or

second as problematical plus ly list takes precedence adjectivesuxally

comfortable is analysed rst as verb sux Verb takes precedence when the

the verb comfort plus the sux or sux is able er or or ure

able and second as the noun noun sux noun takes precedence when the

comfort and the sux able sux is ery ism ist ous

chatty is analysed rst as the some or y

noun chat plus the sux yand

second as the verb chat plus the

sux y

verb sux age When the word denotes action leakage is analysed rst as the

or or an instance of a phenomenon verb leak the sux ageand

noun sux age then the verb takes precedence second as the noun leak the

when the word denotes a measure ax age

or collection of something then

the noun takes precedence

prex a noun Verb takes precedence aglow is analysed rst as the

or prex a plus the verb glow and

prex a verb second as prex a plus the noun

glow

Preferential order for the analysis of comp ounds

verb noun Verb noun takes precedence checkp oint is analysed rst as

or the noun check plus the noun

noun noun p oint and second as the verb

check plus the noun point

noun noun Noun noun takes precedence windfall is analysed rst as the

or noun wind plus the noun fall

noun verb and second as the noun wind

plus the verb fall

Table How to order multiple analyses of comp ounds and derivations

Analysis typ e co des

following this one Then saythatyou are interested in

two stems diveb omber which has three dierent analyses

and typ esetterwhichhastwo Both words are nounverb

ax typ e words whichmaybederivations or comp ounds or

derivational comp ounds and this accounts for four of the

analyses given However for the comp ound analysis of dive

bomber the stem dive is analysed as a verb but can also b e

thoughtofasanounwhichgives an extra analysis

First you can decide whether you want just one default analy

sis for each stem or whether you want to see all the available

analyses

If you want to see all p ossible segmentations then you dont

need to add extra restrictions As the MorphCnt column

indicates there are three analyses given for diveb omber and

twofor typ esetter so this is what the unrestricted example

lexicon lo oks like

Stem MorphNum NVAComp Der Comp DerComp Def Imm ImmClass

divebomber Y Y N N Y divebomber Vx

divebomber Y N Y N Y divebomber VN

divebomber Y N Y N N divebomber NN

typesetter Y Y N N Y typeseter Vx

typesetter Y N N Y Y typeseter NVx

Derivations take precedence over comp ounds so for b oth

words the rst row with analysis number contains the

derivation and gets Y under Der And since eachword

has only one p ossible derivation analysis b oth are also de

fault analyses and therefore get Y under Def to o The

N under Comp and DerComp conrm that they are not

comp ounds or derivational comp ounds

Comp ounds take precedence over derivational comp ounds so

for diveb omber the next tworows contain the two comp ound

analyses with analysis numbers MorphNum and

Both get the co de Y under Comp Since verb noun

comp ounds take precedence over noun noun comp ounds

the verb noun analysis is a default analysis it gets as its

MorphNum and Y under Def The noun noun analysis

gets as its MorphNumand N under DefThe N co des

under Der and DerComp conrm that neither of these

analyses is a derivation or a derivational comp ound

english linguistic guide

The last row in the lexicon gives the derivational comp ound

analysis of typ esetter with Y under DerComp Since it is

the only p ossible derivational comp ound analysis it is also a

default analysis and therefore gets Y under Def to o The

N under Der and Comp conrm that it is not a derivation

or a comp ound

However rather than including all four forms in your lexicon

you mightwant to ignore the derivation and derivational

comp ound analyses and just see the comp ound analyses

To do this for all the stems in the database you should

add an expression restriction to your lexicon which states

that Comp Y In the example lexicon this one restriction

pro duces the following result

Stem MorphNum NVAComp Der Comp DerComp Def Imm ImmClass

divebomber Y N Y N Y divebombe r VN

divebomber Y N Y N N divebombe r NN

In the same wayifyou want to examine derivational com

p ound analyses and leave out all the other analyses you

should add an expression restriction to your lexicon which

states that DerComp Y In the example lexicon this re

striction pro duces the following result

Stem MorphNum NVAComp Der Comp DerComp Def Imm ImmClass

typesette r Y N N Y Y typeseter NVx

Rather than seeing a numb er of analyses you might prefer to

lo ok at just one straightforward default analysis no matter

howmany alternatives are given in subsequentrows Again

you can quickly construct restrictions to make this p ossible

The quickest wayistousetheMorphNum column which

gives a numbertoeach analysis of each stem You can say

MorphNum which means that only the very rst analysis

of each stem app ears in your lexicon

Sometimes there may b e more than one default analysis If

you want to see just the default analysis of each comp ound

you should use these two restrictions Def Y and Comp

Y In the example lexicon this means that the nonpreferred

noun noun analysis is left out

Stem MorphNum NVAComp Der Comp DerComp Def Imm ImmClass

divebomber Y N Y N Y divebombe r VN

Analysis typ e co des

These explanations may app ear complicated but by reading

them you can get to know the imp ortant restrictions that

you can use to extract the typ es of analyses you really want

IMMEDIATE SEGMENTATION

Immediate segmentation is the least detailed form of analysis

oered here It do esnt giveyou a full analysis rightdown

to all the smallest elements a stem contains rather it is a

simple onelevel breakdown of a stem into its next biggest

elements So while complete segmentation is equivalentto

a full analytical tree immediate analysis can b e thoughtof

as a close lo ok at a particular level

There are ten columns which present the immediate segmen

tation of stems to you The rst gives the orthographyofthe

analysed elements The next three give more general co dings

so that using the flex options SHOW and QUERYyou can

lo ok for stems whichhave a particular form a prep osition

plus a noun say or a stem plus a stem plus an ax and

so on The remaining six deal with particular features which

sometimes o ccur in morphological analysis stem allomor

phy ax substitution opacity derivational transformation

inxation and reversion

In the rst column you get the orthography of the rstlevel

elements themselves each separated bya sign Diacrit

ical markers are not included Thus the stem nameplate

is shown as nameplate in accordance with the various

rules discussed in section Note that each elementis

given in the form of a headwordoranaxeven when the

original word do esnt use that particular form Thus the

stem liturgical is analysed as liturgyicalwhere liturg is re

written in the normal form of the stem liturgyTheflex

name and description of this column are as follows

Imm Immediate segmentation

ImmLemma

The second column is like the rst except that where the

rst column gives you the orthographyofeach element this

column gives you the word class of each element leaving out

any signs Single letter lab els are used to representthe

syntactic class of each element whichisunlikemanyofthe

english linguistic guide

syntactic co des used in other parts of the database The use

of a single character means that there is no p ossibilityofa

co de b ecoming ambiguous since eachcharacter is unique

Table shows you the lab els used in this column

Word Class Lab el

Noun N

Adjective A

Numeral Q

Verb V

Article D

Pronoun O

Adverb B

Prep osition P

Conjunction C

Interjection I

Single contraction S

Complex contraction T

Ax x

Table Word class lab els immediate segmentation

Using these co des nameplate is given the co de NN to in

dicate that it is made up of two nouns a comp ound and

emigration has the co de Vx to indicate that it is made up

of a verb and an ax a derivation The flex name and

description of the column that gives you these co des are as

follows

Immediate segmentation word class labels ImmClass

ImmClassLemma

The third column provides more detailed information ab out

the syntactic categorization of verbal stems The basic co des

used are exactly the same as the ImmClass column except

that instead of the V co de to representaverb any one of a

numb er of co des is given Table shows you these co des

along with their meaning

Verbal subcategory Lab el

Intransitive

Transitive

Intransitive transitive

Unmarked for transitivity

Table Onecharacte r verbal sub class lab els

Immediate segmentation

In this column the word emigration has the co de xItis

exactly the same as the co de in the previous column except

that the V is replaced bythenumber indicating in more

detail what sort of verb it is

The flex name and description of this column are as follows

ImmSubCat Immediate segmentation subcat labels

ImmSubCatLemma

The fourth immediate segmentation column simply tells you

whether the elements identied are stems or axes Upp er

case S indicates a stem upp er case A indicates an ax

and upp er case F indicates a ectional form of a stem Thus

emigration is represented as SAandbagpip es as SF The

flex name and description of this column are as follows

ImmSA Immediate segmentation stemaffix labels

ImmSALemma

The fth immediate segmentation column concerns stem al

lomorphy Within a word a stem sometimes takes a form

dierent from the one used when it is written down as a

word in its own right When morphological analysis is noted

down any resulting stems are given their normal stem form

b ecause its easiest to understand An example is the word

abundant which comprises the stem ab ound and the ax

ant Note the dierence b etween what app ears in the original

word abund and its regular stem form ab ound each has

the same meaning the only dierence b etween them is their

sp elling This is an example of derivational stem allomorphy

since a new word has b een derived by linking a dierentform

of stem to an ax

Another sort of stem allomorphy sometimes o ccurs with con

versions that is words whichchange their word class with

out the addition of an ax sleep is b oth a noun and a

verb for example When conversion o ccurs and the form

of the stem seems to have altered the pro cess can b e termed

conversion with allomorphy The verb halve is an instance

of conversion with allomorphy since it is a conversion from

the noun form half Thus half and halve are considered

allomorphs two dierent forms or representations of the

same stem There are three typ es of conversion with allo

morphy The rst is the voicing of the nal consonant with

english linguistic guide

the addition of a nal e thus the verb thieve is considered

to b e a conversion of the noun thief The second is the

same pro cess in reverse the removal of a nal e and the

devoicing of the last consonant thus the noun b elief is a

conversion of the verb believe The third is the change in

sp elling from nal s to c the noun practice can thus b e

thought of as a zeroderivation from the verb practise

The next typ e of stem allomorphyisectional allomorphy

a relatively rare typ e When the irregular past tense of a

verb is used as an adjective b oth are said to derivefromthe

innitive form so that the adjective drunken comes from the

verb drink The same is true for past participle forms the

adjective born thus derives from the verb bear

There are two other categories which are dealt with under

stem allomorphyeven though theyre not really instances of

stem allomorphy clippings and blends Clippings are short

ened forms of words whichdonotchange word class For

example phone is a simple clipping of telephone Sometimes

a clipping consists of more than one morpheme vib es is a

clipping of vibraphone whichcontains the stem vibraphone

and the ax sandhanky is a diminutive form consisting

of the stem handkerchief and the ax y

A blend is a word whichismadeupoftwo stems at least

one of whichmay b e shortened The word smog is made up

of the stems smoke and fogandparatro op er consists of the

stems parachute and tro op er Note that the denition of a

blend only allows for stems not axes

The table b elow summarizes the vetyp es of allomorphy

and shows the co des used to identify them in the ImmAllo

column

Stem Allomorphy Co de Example

Blend B breathalyse

Clipping C phone

Derivational D clarify

Flectional F b orn

Conversion Z b elief

Table Stem allomorphycodes

The flex name and description of this column are as follows

Immediate segmentation

ImmAllo Stem allomorphy top level

ImmAlloLemma

The sixth immediate segmentation column marks stems with

a morphological analysis involving ax substitutionThisis

the pro cess whereby an ax replaces part of a stem when

that stem and the ax join to form another stem For

example active is analysed as the stem action and the ax

ive the ax ion has disapp eared and the new ax ive

has taken the place of the old one So this column gives Y

for yes if the immediate analysis of the stem involves ax

substitution or N for no if it do es not The flex column

name and description of this column are as follows

ImmSubst Affix substitution top level

ImmSubstLemma

The seventh column identies those words whose analysis is

opaque thatiswords made up of morphemes which are

recognisable but where the meaning of the head element

isnt reected in the meaning of the full word An example

of this is accordion it app ears to b e made up of the verbal

stem accord the head element and the ax ion Since

the semantic link b etween accord and accordion is far from

obvious the analysis is marked as b eing opaque and it gets a

Y in this column Words whose analyses are morphologically

and semantically clear get the co de NTheflex name and

description of this column are as follows

Opacity top level ImmOpac

ImmOpacLemma

The eighth immediate segmentation column gives simple ex

pressions to illustrate any orthographic alterations the analy

sis of a word involves A morpheme b oundary is marked by

a and letters removed from either side of a morpheme are

prexed bya and letters which are added are prexed by

a Letters which do not change are considered part of a

morpheme and not shown A simple example is this is

the pattern for the word unable since it consists of the ax

un and the stem able and neither morpheme alters On

the other hand undersized is given as e since nothing

happ ens to the rst morpheme under the nal e of the

english linguistic guide

second morpheme size is removed and nothing happ ens to

the last morpheme edTheflex name and description of

the columns that contain these expressions are as follows

TransDer Derivational transformation top level

TransDerLemma

The ninth column indicates which stems have an immediate

analysis involving derivation by means of an inx Usually

derivational axes are added to the b eginning or end of a

stem but in some cases the ax is inserted into a multi

word as in derivations from verbandparticle combinations

like hangeron from hang on and lo okeron from lo ok on

Stems marked for this typ e of inxation get the co de Y in

this column all other analyses get the co de N

ImmInx Infixation top level

ImmInxLemma

The last immediate segmentation column deals with stems

analysed as conversions from multiwords whichhave under

gone reversion of their parts in the pro cess of conversion For

instance the noun downp our is considered a conversion of

the verb p our down and the adjective oputting is derived

from the verb put o via its ection putting o Whenever

a stem is analysed in this way this column yields a Y co de

In all other cases an N co de is given

ImmRevers Reversion top level

ImmReversLemma

COMPLETE SEGMENTATION FLAT

Complete segmentation is complete in the sense that it iden

ties all the morphemes a stem contains This is in contrast

to immediate segmentation which only picks out the next

two sometimes three morphological elements The com

plete segmentation discussed in this section is also at which

means that you can see what the constituent morphemes are

without knowing the details of the full morphological analysis

which has b een carried out When you draw a morphological

tree diagram this information gives the outermost branches

only you cannot analyse any further and you cannot see the

Complete segmentatio n at

intermediate levels So when you want to see the complete

at segmentation of counterrevolutionary for example you

get this sort of information

counterrevolutionary

counter revolt ution ary

There are three columns with complete segmentation at

information The rst contains the morphemes themselves

The second contains the word class of each morpheme and

the third simply states whether each morpheme is a stem or

an ax The last two columns are useful when youre lo ok

ing for a stem with a particular combination of morphemes

using the flex SHOW and QUERY options you can huntout

stems which are made up of a noun plus an ax plus a noun

say or all the stems which contain at least three other stems

The rst column gives you each stem split into its morphemes

by signs Thus the stem counterrevolutionary is written

in the following way

counterrevoltutionary

No diacritics are included The flex name and description

of this column are as follows

Flat Flat segmentation

FlatLemma

The second column uses singleletter co des to representthe

word class of each morpheme

english linguistic guide

Word Class Lab el

Noun N

Adjective A

Numeral Q

Verb V

Article D

Pronoun O

Adverb B

Prep osition P

Conjunction C

Interjection I

Single contraction S

Complex contraction T

Ax x

Table Word class lab els at segmentation

Using these co des the stem counterrevolutionary is given

as xVxxTheflex name and description of the column are

as follows

Flat segmentation word class labels FlatClass

FlatClassLemma

The last column simply indicates whether each morpheme

is a stem a ection or an ax Upp er case S means Stem

upp er case F means Flection and upp er case A means Ax

The full co de for counterrevolutionary is thus ASAA The

flex name and description of this column are as follows

FlatSA Flat segmentation stemaffix labels

FlatSALemma

COMPLETE SEGMENTATION HIERARCHICAL

Complete hierarchical segmentation gives the most detailed

analysis available for each stem It is called hierarchical

b ecause it can cover several dierentlevels it is arrived

at after immediate analysis has b een carried out on every

stem that can b e identied within a larger stem With

this information you can draw a complete morphological

tree diagram from the ro ot to the outermost branches

with every intermediate branch fully represented So for

Complete segmentation hierarchical

the stem counterrevolutionaryyou can get the following

morphological analysis

counterrevolutionary

counterrevolu tion

revolution

counter revolt ution ary

There are six columns which give information ab out the full

segmentations of stems Three of them give the hierarchical

segmentations themselves The simplest of these tells you

what the constituent morphemes of the stem are indicating

with algebralikebrackets the structure of the tree Also

available are similar bracket notations which supply a word

class lab el alongside each element on eachlevel or the word

class without the sp elling of the element itself The remain

ing two columns indicate whether stem allomorphyorax

substitution has o ccurred anywhere in the full hierarchical

analysis

The rst column provides all the information you need to

draw a tree diagram like the one ab ove that is the con

stituent morphemes of a stem each delimited by a comma

and enclosed in brackets which indicate its complete mor

phological structure The stem counterrevolutionary thus

lo oks like this

counterrevoltutionary

Each identiable stem or ax is enclosed by a pair of brack

ets b eginning with the brackets round the full original stem

Then there are brackets round the stem counterrevolution

and subsequently round the stem revolution Finally there

are brackets round each of the four morphemes

The flex name and description of the column which contains

morphological analyses in this form are as follows

Struc Structured segmentation StrucLemma

english linguistic guide

The next two columns use extra lab els to indicate the word

class of each segment They are given b etween square brack

ets to the rightofeach closing round bracket so that every

segmentonevery level within the original stem has a word

class co de The word class co des used are as follows

Word Class Lab el

Noun N

Adjective A

Numeral Q

Verb V

Article D

Pronoun O

Adverb B

Prep osition P

Conjunction C

Interjection I

Single contraction S

Complex contraction T

Table Word class lab els complete segmentation

The co des used for axes are combinations of these word

class lab els The stem counterrevolutionary can b e repre

sented as follows

counterNNrevoltVutionNVNN aryANN

This example illustrates the sp ecial form ax co des take

There are two elements in each ax co de which are separated

byavertical bar Infront of the vertical bar is a single co de

which is the word class of the stem which the ax in question

helps to form After the vertical bar comes a combination

of single letter co des which indicate the word class of each

element within the stem formed and the p osition of the ax

itself is given by a dot

In the counterrevolutionary example ab ove the co de given

alongside the ax counter is NN The N b efore the bar

means that the ax counter helps to form a stem whichis

a noun counterrevolution The N after the bar means

that the segmentation of the noun counterrevolution is ax

plus noun These detailed co des can help you to identify the

way axes are used and to get lists of stems whichcontain

axes used in particular contexts the fact that the second

part of the counter co de is N helps you to see at once that

Complete segmentation hierarchical

this ax helps to form a derivation in conjunction with a

noun

Sometimes a pair of axes can only b e used together as in

the word aero drome theword aero do es not exist and the

word drome do es not exist In such cases x marks the other

ax and denotes that the axes must o ccur in combination

with each other socalled combining forms Thecodefor

the aero of aero drome is thus Nxandthecodeforthe

drome is Nx

So this column is particularly useful for two things First

you get the word class of each stem in the segmentation

alongside the orthographic representations of individual mor

phemes Second you get detailed information ab out each

ax each stem contains The flex name and description of

this column are as follows

Structured segmentation word class labels StrucLab

StrucLabLemma

The next column shows the hierarchical structure of each

stem by means of round brackets and commas and the full

word class lab els b etween square brackets just as with the

previous column The only dierence is that in this column

the orthographic representation of the constituent stems and

axes is missed out altogether Thus the stem counter

revolutionary gets the following representation

NNVNVNNANN

This column again helps you to search for stems whichhave

a particular morphological structure and particular combina

tions of syntactic elements The flex name and description

of this column are as follows

StrucBrackLab Structured segmentation word class labels only

StrucBrackLabLemma

The fourth hierarchical segmentation column deals with stem

allomorphy Within words stems sometimes take a form

dierent from their generally accepted stem form When a

morphological analysis is noted down the resulting stems

are given their normal stem orthography An example is the

word inediblewhich comprises the ax in the stem eat

english linguistic guide

and the ax ible note the dierence b etween ed and eat

where the one elementisspelttwo dierentways This is

stem allomorphy If stem allomorphy o ccurs at anypointin

a stems complete hierarchical segmentation a co de is given

in this column to show what sort of stem allomorphy o ccurs

The table b elowshows the co des and you can read more

ab out what each co de means in section ab ovethey

are the same co des used in the ImmAllo column

Stem Allomorphy Co de Example

Blend B breathalyse

Clipping C phone

Derivational D clarify

Flectional F b orn

Conversion Z b elief

Table Stem allomorphycodes

The flex name and description for this column are as fol

lows

StrucAllo Stem allomorphy any level

StrucAlloLemma

The fth hierarchical segmentation column marks stems with

a morphological analysis involving ax substitutionThisis

the pro cess whereby an ax replaces part of a stem when

that stem and the ax join to form another stem For

example melo dic is analysed as the stem melo dy plus the

ax ic the ax y has disapp eared and the new ax ic

has taken the place of the old one So this column gives

Y for yes if the complete analysis of the stem involves ax

substitution or N for no if it do es not The flex name and

description of this column are as follows

StrucSubst Affix substitution any level

StrucSubstLemma

The sixth and last hierarchical segmentation column identi

es those words whose analysis is completely or partly opaque

that is words made up of morphemes which are recog

nisable but where the meaning of the head element isnt

reected in the meaning of the full word An example of

this is ladykiller it app ears to b e made up of the noun

Complete segmentation hierarchical

stem lady and the noun stem killer which can subsequently

b e analysed as kill plus er Since the meaning of the

head element killer do esnt relate directly to the meaning

of the full word the analysis is marked as b eing opaque

and it gets a Y in this column Words whose analyses are

morphologically and semantically clear get the co de N The

flex name and description of this column are as follows

StrucOpac Opacity any level

StrucOpacLemma

OTHER CODES

The remaining three columns give counts of various sorts

the number of comp onents ie stems and axes in the

immediate analysis of each stem the number of morphemes

astemcontains after complete segmentation and the number

of levels involved in the complete hierarchical analysis of each

stem

The rst of these columns is the simple count of the number

of comp onents each stem contains The normal gure is

two words are generally split into two parts each time one

level of morphological analysis takes place Sometimes three

comp onents can b e identied derivational comp ounds are

usually analysed as a stem plus a stem plus an ax as

are normal comp ounds which are joined with a sp ecial link

morpheme a oor s And of course monomorphemic

words only contain one comp onent Any stems which cannot

receive an adequate morphological analysis for the reasons

given in section get the number

Some examples in the stem counterrevolutionary the num

b er of comp onents is twothestemcounterrevolution and

the ax ary and for lawbreaker it is three the stem law

the stem break and the ax er

The flex name and description of this column are as follows

CompCnt Number of morphological components

CompCntLemma

english linguistic guide

The second column gives you the numb er of morphemes in

each stem For words without a morphological analysis the

numb er given is zero The numb er of morphemes in the

stem counterrevolutionary for example is four while for law

breaker it is three

The flex name and description of this column are as follows

MorCnt Number of morphemes

MorCntLemma

The last of the three columns gives a count of the number

of levels in the complete hierarchical segmentation describ ed

ab ove which is b est illustrated by another lo ok at the tree

diagram illustrating the analysis of counterrevolutionary

counterrevolutionary

counterrevolu tion

revolution

counter revolt ution ary

Including the stem at the top the diagram covers four lines

this is the number of levels the stem has It is the number

of times you can carry on doing immediate analysis when

you analyse a particular stem in full Do not confuse it with

the numb er of all the immediate analyses required to arrive

at the complete hierarchical segmentation anyonelevel of

analysis may include more than one immediate segmentation

Monomorphemic stems always get the numb er while stems

without analysis for reasons explained in section get

the number

The flex column name and description of this column are

as follows

LevelCnt Number of morphological levels

LevelCntLemma

Morphology of English wordforms

MORPHOLOGY OF ENGLISH WORDFORMS

There are twotyp es of morphology information available

for the wordforms given in the celex database rst in

formation ab out the lemma which underlies each family of

wordforms and second a simple identication of the inec

tional features which are sp ecic to eachwordform either in

the form of thirteen yesno feature columns or one column

with feature identication co des

Dictionaries present their lexical information under b old

typ e headwords which are used instead of listing every indi

vidual inected form separatelySuch a form is often called

the canonical form since it represents a full canon of inec

tions Thus the word eat is understo o d as referring not only

to the form eat itself but also the forms eats eating ate

and eatenToprint full details ab out every inected form

separately would result in a lot of needless rep etition and

enormous b o oks which no one could lift from the b o okshelf

However for many applications lemma information has to b e

listed for each individual wordform and in a celex lexicon

of typ e wordform you can do just that when you include

certain morphological columns This is done byproviding a

link b etween the wordform information and the lemma infor

mation When you cho ose the option Lemma information

from the ADD COLUMNS menu you are in fact b eing allowed

into the lemma information by the backdoorYou can now

lo ok up information sp ecic to a particular wordform in your

lexicon and at the same time see general information which

is common to all the other forms in the same inectional

paradigm One particularly useful typ e of lemma information

you can use in your wordform lexicon is the syntactic in

formation which can givethewordclassofanywordform

you are lo oking at There is also an imp ortant distinction

whichyou may b e able to draw up on with the frequency

information The wordform lexicon gives you a cobuild

frequency gure sp ecic to eachwordform while the lemma

information available lets you see the sum frequency for all

the inectional forms in the same paradigm a gure referred

to as the lemma frequency

All the lemma information has already b een dened else

where in this linguistic guide so there is no p ointinrepeat

ing it here All that needs to b e p ointed out is that the column names used in a real lemma lexicon dier from those

english linguistic guide

used in the lemma information option in the morphology of

wordforms When a flex column name and description are

dened in the course of lemma lexicon text the column name

given in brackets is the name of the column when it is used

as part of a wordforms lexicon Usually this name is identical

to the lemma lexicon name except that the word lemma is

addedtotheend

ExampleName The column names used for lemma information

ExampleNameLemma in a Wordforms lexicon are given in

brackets as this Example Name shows

All the other details and denitions remain the same in b oth

cases So when youre lo oking for the columns of lemma

information provided with a wordforms lexicon under mor

phologyjustgoback to the original lemma information its

all there

INFLECTIONAL FEATURES

There are thirteen sp ecial columns available only with a lexi

con of typ e wordforms Each one corresp onds to a particular

inectional attribute whicha wordform can have There can

only b e one of two co des in each column Y for yes this

wordform has this attribute or N for no this wordform do es

not have this attribute These columns are therefore useful

for constructing restrictions on your lexicons restrictions

which need not b e on view its unlikely that you will

wanttolookatthecontents of these columns with the SHOW

option If on the other hand you wanttohave a lab el

whichletsyou see at a glance all the inectional features

eachwordform has then you should use the typ e of ection

co des describ ed in the next section

An example Tomake a lexicon which gives you all rst

p erson present tense verb forms in the database you have

to include at least three columns in the wordforms lexicon

you create namely a column which gives the orthographic

representations you prefer along with Pres and Sin which

are amongst the thirteen columns describ ed b elow You

must then construct two restrictions for your lexicon one

stating that Pres must b e equal to Y and another stat

ing that Sin mustbeequalto YYou can then format

Inectional features

your lexicon to make sure that Pres and Sin are not on

view that waywhenyou SHOW or EXPORT your lexicon

you just get the list of words you require without two lists

of Ys To this basic lexicon you can of course add any

other columns you require either the orthographic and fre

quency information sp ecic to eachwordform or the general

lemma informationparticularly syntaxwhichisavailable

through the Morphology of English wordforms options

The rst inectional features column indicates whether a

wordform is a singular form of anysort This means past

and present tense verb forms suchashib ernated or babbles

or nouns suchassagacityTheflex name and description

of this column are as follows

Sing Inflectional feature singular

The second column indicates whether a wordform is a plural

inection of any sort This means past and present tense

verb forms suchas hib ernate or submerged or nouns suchas

jo cularitiesTheflex column name and description of this

column are as follows

Plu Inflectional feature plural

The third column marks all the wordforms which are p osi

tive forms that is not comparative or sup erlative forms

like b etter and best but plain adjectival forms like go o d or

oftenThus adjectives like go ofy and idiomatic or adverbs

like seldom and idiomatically get the co de Y while all other

forms get the co de NTheflex name and description of this

column are as follows

Pos Inflectional feature positive

The fourth column marks all the wordforms which are com

parative forms almost always adjectives Wordforms such

as b etter or angrier or cannier thus get the co de Y while

all other noncomparative forms get the co de N There is

also a small numb er of comparativeadverbs which get the

Y co de suchasfurther The flex name and description of

this column are as follows

Comp Inflectional feature comparative

english linguistic guide

The fth column marks all sup erlative forms so that word

forms suchasbest or angriest get the co de Y and every

other form gets the co de N There is also a small number of

sup erlative adverbs which get the Y co de suchasfurthest

The flex name and description of this column are as follows

Sup Inflectional feature superlative

The sixth column marks the form of the verb usually known

as the innitive It is used as a headword in the celex

databases and in most dictionaries Words like wae or

have or eat which can b e used with the particle to in front

of them are innitives Anywordform which is an innitive

gets a Y co de in this column all the others get the co de N

The flex column name and description for this column are

as follows

Inf Inflectional feature infinitive

The seventh column marks any participles past tense or

present tense Present participles are normally formed by

adding ing to the stem of the verb with the exception of

some irregular verbs Past participles add a sux ending

in d to the stem and they are used in the formation of

the p erfect tense Ive lived in Nijmegen for four years

Again many irregular verbs dont match this rule gone is

the past participle of go for example Most past participles

can also b e used adjectivallyasinthe panelled walls Any

wordforms which are participles get the co de Y and all the

rest get the co de NTheflex name and description of this

column are as follows

Part Inflectional feature participle

The eighth column identies any present tense forms in

cluding the present participles mentioned under PartThus

verb forms like gleam gleams and gleaming get the co de Y

while all other forms including innitives which are marked

in a dierent column get the co de NTheflex name and

description of this column are as follows

Pres Inflectional feature present tense

Inectional features

The ninth column identies any past tense forms including

the past participles mentioned under Part Thus forms

like o ccupied and elicited get the co de Y while all other

forms including innitives whicharemarked in a dierent

column get the co de N The flex name and description of

this column are as follows

Past Inflectional feature past tense

The tenth column marks rst p erson singular forms of verbs

whether present tense or past tense So all rst p erson

singular forms likeI go orInished o are given the

co de Y and every other form gets the co de NTheflex

column name and description of this column are as follows

Sin Inflectional feature st person verb

The eleventh column marks second p erson singular forms of

verbs whether present tense or past tense forms For most

verbs the second p erson form is the same as the rst p erson

form but some irregular verbs are exceptions So all second

p erson forms likeyou are oryou shout aregiven the co de

Yandevery other form gets the co de NTheflex column

name and description of this column are as follows

Sin Inflectional feature nd person verb

The twelfth column identies third p erson singular forms of

verbs whether present tense or present tense forms For

most verbs the third p erson present tense form consists of

the stem plus the sux sThus forms likehe sto o d up or

Gilb ert acts get the co de Y while every other form gets the

co de N The flex name and description for this column are

as follows

Sin Inflectional feature rd person verb

The thirteenth and last column marks rare forms normally

forms whichhave b ecome outdated like brethren shouldst

or wertSuch forms have the co de Y in this column while

every other wordform gets the co de N The flex name and

description of this column are as follows

Rare Inflectional feature Rare form

english linguistic guide

TYPE OF FLECTION

In the Inectional Features section ab ove thirteen dierent

inectional features are distinguished and assigned to thir

teen separate yesno columns The same information is also

available in one single column using combinations of single

letter co des to show all the features eachwordform has The

yesno columns are useful for constructing restrictions on

your lexicon whereas the typ e of ection column describ ed

here provides you with a lab el that identies at a glance all

the features eachwordform has Table b elow sets out all

the combinations of singleletter co des that o ccur

Inectional feature Lab el yesno column name

Singular S Sing

Plural P Plu

Positive b Pos

Comparative c Comp

Sup erlative s Sup

Innitive i Inf

Participle p Part

Presenttense e Pres

Past tense a Past

st p erson verb Sin

nd p erson verb Sin

rd p erson verb Sin

Rare form r Rare

Headword form X

not nouns verbs

adjectives or adverbs

Table Typ e of ection lab els

For a full denition of these ection typ es read the details

given for the appropriate yesno columns in section

ab ove However note that there is one typ e of ection lab el

which do es not corresp ond to a yesno column The X

lab el identies many forms not covered by the other lab els

including prep ositions like among or less pronouns like that

or hers conjunctions like immediately or thatnumerals like

fth or thousandcontracted forms like Ill or hadntand

interjections like phew or amen These forms are always the

same as those used as the headword form of the lemma thus

the very few inected adverbial forms do not get the co de

Typ e of ection

X No nouns verbs adjectives or adverbs ever get the co de

X

Eachwordform mayhave more than one co de attached to it

Thus the wordform b oasted has the co de aS a means it

is a past tense form means that it is a third p erson form

and S means that it is singular

The flex name and description of this column are as follows

FlectTyp e Type of flection

INFLECTIONAL TRANSFORMATION

The last column shows how the orthographic form of a stem

is altered when a ection is formed Each string of letters in

the stem is shown by the symbol so the rst p erson present

tense form of the verb whose stem is abide is simply given as

Any blanks or hyphens in the stem are shown as a blank

so abide by is shown as Letters removed from the front

or back of a string are prexed by a minus sign and letters

added are prexed by a plus sign soabiding by is given as

eing rst the nal e of abide is removed and then

the sux ing is added This formalism is an unambiguous

wayofshowing the inectional transformations that o ccur in

the orthographic formation of wordforms

Whenever the inectional transformation is irregular as with

the past tense forms of the verb sing for example sang and

sung no transformation is given the eld remains empty

The tables b elowshow all the lettergroups represented in

the database which can b e subtracted from or added to a

headword to makeawordform

Lettergroups removed from a headword

e ey f fe y

Table Inectional transformat ion co des letters removed

english linguistic guide

Lettergroups added to a headword

bed ber best bing d

ded der dest ding ed

er es est ged ger

gest ging ied ier ies

iest ing ked king led

ler lest ling med mer

mest ming ned ner nest

ning ped ping r red

ring s sed ses sing

st ted ter test ting

ved ving ves zed zes

zing

Table Inectional transformat ion co des letters added

The flex name and description of this column are as follows

TransIn Inflectional transformation

TransInLemma

English Syntax

ENGLISH SYNTAX

Syntactic information is available for lexicons of typ e lemma

or with the lemma information presented with Morphology of

English Wordforms The subsections whichmake up this sec

tion corresp ond to the seven sub classication options flex

oers you when you ask for syntactic information under the

add columns window

ADD COLUMNS

Word Class

Subclassificatio n nouns

Subclassificatio n verbs

Subclassificatio n adjectives

Subclassificatio n adverbs

Subclassificatio n numerals

TOP MENU

PREVIOUS MENU

v

ADD COLUMNS

Subclassificatio n pronouns

Subclassificatio n conjunctions

TOP MENU

PREVIOUS MENU

First and foremost as shown in the two add columns

menus there are basic word class co des available for all the

lemmas in the database in the form of numb ers or lab els

Then there is a wealth of sub classication information which

supplements the basic word class co des for eachofthe

more imp ortant classes of lemma there are a number of

columns which indicate whether or not a certain lemma has

english linguistic guide

a particular feature For example you can quickly see that

the verb lch is transitive and the verb zzle is not when you

check the entries for these words in the Trans V column

An explanation of the format used for the sub classication

columns is given in section

WORD CLASS CODES LETTERS OR NUMBERS

There are twoways of representing the syntactic co de of each

lemma You can cho ose for yourself whether to use numb ers

Numeric co des or shortened verbal co des Lab els An

adverb for example is represented by the number or the

letters ADV No matter whichtypeofcodeyou decide to use

the information remains the same only the format changes

Examples are provided in the table b elow

Word class co des are a simple way of identifying the syntactic

class of every lemma in the database In addition they

also identify the word class of every wordform since all the

wordforms whichmake up the inectional paradigm of any

given lemma naturally share the same word class So if you

wanta wordform lexicon which gives you the word class

of anywordform you have to use the lemma information

columns whichareavailable under Morphology of English

Wordforms

Twelve basic categoriesset out in Table b eloware dis

tinguished and you can use them in either of the two forms

describ ed here Only the last two classications given may

need some explanation they deal with contractions those

words which are made up of shortened forms of other words

For example isnt is a contraction of is and notanddyou

is a contraction of do and you Twotyp es of contraction

are identied A simple contraction is a shortened form

in isolation ve and re are b oth simple contractions A

complex contraction is one form which contains twowords

one of which might b e shortened Ive and couldnt are b oth

complex contractions

Word class co des letters or numb ers

Word Columns Example

Class

ClassNum Class

Noun N garrison

Adjective A biblical

Numeral NUM twentieth

Verb V chortle

Article ART the

Pronoun PRON mine

Adverb ADV sheepishly

Prep osition PREP through

Conjunction C whereas

Interjection I alleluia

Single contraction SCON re

Complex contraction CCON youre

Table English word class co des

If you wantword class co des in the form of numb ers then

cho ose the column which has the following flex name and

description

Word class numeric ClassNum

ClassNumLemma

If you wantword class co des in the form of short verbal

symb ols cho ose the column which has the following flex

name and description

Word class labels Class

ClassLemma

SUBCLASSIFICATION Y OR N

All the remaining syntactic columns come under a general

heading of sub classications Each of these sub classication

columns represents a particular syntactic attribute that a

lemma mighthave The values contained in each column can

only b e one of twocharacters Y or N Using these columns

you can carry out quickchecks on the qualities of any lemma

Is the conjunction whereas a co ordinating conjunction Lo ok

up its value in Cor C and you discover that N no it is not

Is the noun brother avo cative form Check its value in the

column Voc N and you discover that Yyes it is or rather

yes it can b e used in this way

english linguistic guide

Of course its p ossible to do more with these columns than

simple checks on individual lemmas You can use them to de

ne the contents of your lexicon by constructing a restriction

with one or more of them If you wantsay a lexicon which

only contains entries that can b e ditransitiveverbs then you

should rst include the column Ditrans V in your lexicon

then construct a restriction which states that Ditrans V

Y Moreover you can construct anynumb er of restrictions in

this way to make the contents of your lexicon even more

closely dened

Rememb er though that these Y N answers are not exclu

siveanswers The verb write can b e b oth transitiveShe

writes very long letters andintransitiveWhat do es he do

for a living He writes So under the columns Trans V

and Intrans V write gets the co de YItfollows that if you

wantalistofverbs whichare always intransitive you must

V construct two restrictions which state rst that Trans

N and second that Intrans VYTo get the b est from

these sub classication co des then you should read the de

scriptions given b elow carefullywork out exactly howtoget

what you want using restrictions and then use flex to build

carefully planned lexicons

SUBCLASSIFICATION NOUNS

There are eleven p ossible attributes which a noun can have

so two add columns menus are needed to cover them all

The v and symb ols in the b ottom right hand corner

of either screen indicate that you should use the next and

prev keys to move from one part of the menu to the other

ADD COLUMNS

Count

Uncount

Singular use

Plural use

Group Count

Group Uncount

TOP MENU

PREVIOUS MENU v

Sub classicat ion nouns

ADD COLUMNS

Attributive

Postpositive

Vocative

Proper noun

Expression

TOP MENU

PREVIOUS MENU

What follows now is a concise description of each of these ten

attributes along with the column names and descriptions

used in flex

The rst column answers the question is this lemma a count

noun Such nouns can b e treated as individual units and

thus counted you can talk ab out seventysix tromb onesbut

not twomusicsThus the noun lemma trombone gets the

co de Y b ecause it is a count noun while the noun lemma

music gets the co de N becauseitisntacount noun And of

course every lemma with a word class other than noun auto

matically gets the co de N The flex name and description

of this column are as follows

C N For nouns countable

C NLemma

The second column a mirror image of the rst it answers the

question is this lemma an uncountable noun Uncountable

nouns are those nouns which are continuous entities they

are not individual units which can o ccur in b oth singular and

plural forms You cannot talk ab out a dozen handwritings

but it is quite p ermissible to talk ab out twelve p ens The

lemma handwriting thus gets the co de Y while the lemma

pen gets the co de N Another term used for uncountable noun

is mass noun

Note however that some nouns app ear to b e b oth countable

and uncountable The noun space is an uncountable noun

when used in a phrase like Space the Final Frontierbut

countable when used in a phrase like The Netherlands at

english linguistic guide

op en spaces This is one example of a lemma whichgetsa Y

in the count and the uncount column

Any noun lemma which can b e uncountable gets the co de Y

in this column all other lemmas get the co de NTheflex

name and description of this column are as follows

Unc N For nouns uncountable

Unc NLemma

The third column answers the question do es this lemma only

ever o ccur in the singular form This refers to what are

sometimes called singularia tantum The lemma monop oly

when used in a phrase like they think they have a monop oly

on the truth can only b e singular and it therefore gets the

co de Y in this column Note however that in other cir

cumstances monop oly do es o ccur in the plural the phrase

private or state monop olies for example Any noun lemma

that can o ccur in a singularonly form gets the co de Y in this

column all other lemmas get the co de N The flex name

and description of this column are as follows

Sing For nouns singular use N

Sing NLemma

The fourth column answers the question do es this lemma

ever o ccur in a pluralonly form This refers to what are

often called pluralia tantum The noun lemma p eople o ccurs

in electiontime phrases like the p eople havespoken where it

means the general p opulation of a country or region It only

o ccurs with a plural verb form and therefore gets the co de Y

in this column Any other lemmas which can never b e plural

only nouns get the co de N The flex name and description

for this column are as follows

Plu N For nouns plural use

Plu NLemma

The fth and sixth column deal with collective nouns those

nouns which refer to groups of p eople or things the lemma

ma jority for example Usually all such lemmas can takea

plural or a singular form of a verb the ma jorityisinfavour

is as acceptable as the ma jority are in favourHowever only in certain cases do es the inectional paradigm of the lemma

Sub classicat ion nouns

include a plural form The lemma mankind for example

can take a plural or a singular verb but it has no plural

form mankinds In contrast the lemma crew can takea

plural or a singular form of the verb and it has a plural form

crews

The fth column answers the question is this lemma a collec

tive noun that has a singular and a plural form The lemma

government gets the co de Y b ecause its always p ossible

to talk ab out a particular say the French government as

well as several say the Europ ean governments The lemma

mankind gets the co de N since it never has a plural form

under any circumstances Other lemmas which are neither

collective nor nouns also get the co de NTheflex name and

description of this column are as follows

N For nouns group countable GrC

GrC NLemma

The sixth column answers the question Is this lemma a

collective noun that only has a singular form and not a

plural form The lemma p opulace has the co de Y b ecause

it is never p ossible to use the form p opulaces and for the

same reason the lemma mankind also gets a Y co de The

flex name and description of this column are as follows

GrUnc For nouns group uncountable N

GrUnc NLemma

The seventh column answers the question can this lemma b e

used attributively This refers to nouns like machine which

o ccurs in phrases like machine translation the rst word says

something ab out the word that follows it So in this column

machine gets the co de Y since it can b e used attributively

A lemma like gadgetry on the other hand gets the co de N

b ecause it cannot b e used attributivelyTheflex name and

description of this column are as follows

Attr N For nouns attributive

Attr NLemma

The eighth column answers the question can this lemma ever

b e used in a p ostp ositiveway Here p ostp ositive refers to

nouns which come after another noun and which qualify that

english linguistic guide

other noun An example is pro of which o ccurs in phrases

like Bushmills whiskey is fortytwo p ercentproof The flex

name and description of this column are as follows

N PostPos For nouns postpositive

PostPos NLemma

The ninth column answers the question is this lemma used

to address p eople or things This usually refers to nouns like

chickenwhich can b e used when you are sp eaking directly

to a hen as in the wellknown song chickchickchickchick

chicken lay a little egg for me or when you are sp eaking

to someone you consider to b e a coward Chicken Come

back and ght The noun chicken thus gets the co de Y

b ecause it can b e used as a vocative whereas a noun dao dil

gets the co de N b ecause normally the more owery p o ets

excepted it cannot b e used as a vo cative The flex name

and description of this column are as follows

Voc N For nouns vocative

Voc NLemma

The tenth column answers the question is this lemma used

as a prop er noun Prop er nouns are the names of p eople or

places so that lemmas suchasArthur or York get the co de Y

while lemmas which arent prop er nouns get the co de N Most

of the prop er nouns in the database are included b ecause they

form a necessary part of the morphological analyses provided

in the Morphology of English lemmas columns The flex

name and description of this column are as follows

N For nouns proper noun Prop er

Prop er NLemma

The eleventh and last column answers the question is this

noun lemma only ever used in combination with certain other

words to make up a particular phrase An example is the

word loggerheads whichonlyever o ccurs in the phrase at

loggerheadsand brunt whichonlyever o ccurs in the phrase

b ear the bruntofSuch nouns get the co de Y while all other

words get the co de NTheflex name and description of this

column are as follows

N For nouns expression Exp

Exp NLemma

Sub classicat ion nouns

SUBCLASSIFICATION VERBS

There are nine verbal sub classication columns available

and they all take the form of Y N answers to questions as

describ ed in section ab ove They are presentedtoyou

in two add columns windows

ADD COLUMNS

Transitive

Transitive plus complementation

Intransitive

Ditransitive

Linking verb

Phrasal

TOP MENU

PREVIOUS MENU

v

ADD COLUMNS

Prepositional

Phrasal prepositional

Expression

TOP MENU

PREVIOUS MENU

The rst of the nine columns answers the question is this

averb which can sometimes take a direct ob ject The

verb lemma crash for example gets the co de Y b ecause you

can say things like he crashed the car So do es the verb

admit b ecause you can say things like he admitted that he

was wrong where the direct ob ject is a clause The verb

lemma cycle gets the co de N because you cant say things

like he cycled the bikeTheflex name and description of

this column are as follows

Trans V For verbs transitive

Trans VLemma

english linguistic guide

The second column answers the question is this a verb which

has an ob ject complement In a phrase like the jury found

him guilty there is a direct ob ject him plus a complement

which relates to that direct ob ject guilty These ob ject com

plements can take the form of a noun phrase they had made

him chairman or an adjective phrase so he thoughtitodd

or a prep ositional phrase when they threw him into jail or

an adverb phrase and kept him there or a clause since it

caused him to b e embarrassed Verbal lemmas which can

takesuch complements get the co de Y in this column all

other lemmas get the co de N The flex name and description

of this column is as follows

TransComp V For verbs transitive plus complementation

VLemma TransComp

The third verbal sub classication column answers the ques

tion is this a verb which sometimes cannot take a direct

ob ject The verb alight for example can never take a direct

ob ject he got the bus and alighted at the CityHall The

verb leave can o ccur with or without a direct ob ject she

left a will or she left at ten oclo ck Both verbs thus get the

co de Y in this column The verb mo dify gets the co de N

however since it always takes a direct ob ject you cannot

say he mo died butyou can say he mo died his opinion on

the matterTheflex name and description of this column

are as follows

V Intrans For verbs intransitive

Intrans VLemma

The fourth column answers the question is this a verb which

can b e ditransitive Here ditransitive refers to verbs which

can taketwo ob jects one direct ob ject plus one indirect

ob ject The verb envy for example gets the co de Y in

this column since you can say he envied his colleagues their

success So do es tell b ecause you can say things like she

told him she would keep in touchVerbs like danceonthe

other hand which cannot taketwo ob jects and all other

nonverbal lemmas get the co de N The flex name and

description of this column are as follows

Ditrans V For verbs ditransitive

Ditrans VLemma

Sub classica tion verbs

The fth verbal sub classication column answers the ques

tion is this lemma ever a linking verb The copula be is a

linking verb it links a sub ject I with a complement which

describ es that sub ject a do ctor in a sentence like Iama

do ctor These sub ject complements can take the form of

a noun phrase she is an intelligentwoman an adjective

phrase she lo oks worried a prep ositional phrase she lives

in Cork an adverb phrase how did she end up there

oraclauseher main intention is to move somewhere else

Verbs which can havesuch sub ject complements get the co de

Y in this column all other lemmas get the co de NTheflex

name and description of this column are as follows

Link V For verbs linking verb

Link VLemma

The sixth column answers the question is this lemma a

phrasal verb A phrasal verbisaverb whichislinked to a

particular adverb suchassp eak out or run away Often these

combinations acquire a sp ecic idiomatic meaning Phrasal

verbs in the database get the co de Y if they are marked

as suchinvolume one of the Oxford Dictionary of Current

Idiomatic English All other lemmas get the co de N The

flex name and description of this column are as follows

Phr V For verbs phrasal verb

VLemma Phr

The seventh column answers the question is this lemma a

prep ositional verb A prep ositional verb is one whichis

linked to a particular prep osition suchasminister to or

consist of Prep ositional verbs in the database get the co de

Y if they are marked as suchintheOxford Dictionary of

Current Idiomatic English All other lemmas get the co de

NTheflex name and description of this column are as

follows

Prep V For verbs prepositional verb

Prep VLemma

The eighth column answers the question is this lemma a

phrasal prep ositional verb A phrasal prep ositional verb is

naturally one whichislinked to a particular adverb and also

to a particular prep osition suchaswalk awaywith or cry

english linguistic guide

out against Phrasal prep ositional verbs in the database are

current phrasal prep ositional verbs only those which are in

use nowadays are given on the basis of their inclusion in the

Oxford Dictionary of Current Idiomatic English All phrasal

prep ositional verbs get the co de Y and all other lemmas get

the co de NTheflex name and description of this column

are as follows

PhrPrep V For verbs phrasal prepositional verb

PhrPrep VLemma

The ninth and last column answers the question is this verb

lemma only ever used in combination with certain other

words to make up a particular phrase An example is the

verb to ewhichonlyever o ccurs in the phrase to e the line

and bell whichonlyever o ccurs in the phrase b ell the cat

Suchverbs get the co de Y while all other words get the

co de N The flex name and description of this column are

as follows

Exp V For verbs expression

Exp VLemma

SUBCLASSIFICATION ADJECTIVES

There are ve attributes of adjectives covered in this version

of the database which are shown in the add columns

windowbelow Each of these attributes or sub classications

has its own column in the database and it always contains

the co de Y or N this simple co ding system is explained

in section ab ove Each of the ve attributes and their

database columns are explained b elow

ADD COLUMNS

Ordinary

Attributive

Predicative

Postpositive

Expression

TOP MENU PREVIOUS MENU

Sub classicat ion adjectives

The rst adjective sub classication column is a simple one it

answers the question is this lemma an ordinary adjective

where ordinary means that it can b e used b oth attributively

the new b o ok and predicatively the b o ok is new Thus

adjectives suchasnew and elementary get the co de Y be

cause they are ordinary adjectives while actual and ablaze

get the co de N b ecause you cannot say the reason is actual

or the ablaze house The flex name and description of

this column are as follows

Ord A For adjectives ordinary

Ord ALemma

The second column gives an answer to the question is this

lemma an adjective whichinsomecontexts can only b e

used attributively Here attributive means those adjectives

which always come b efore the noun or phrase as in sheer

nonsense where sheer cannot come after the noun it qualies

The flex name and description of this column are as follows

Attr For adjectives attributive A

ALemma Attr

The third column answers the question is this lemma an

adjective which in some contexts can only b e used predica

tively Here predicatively refers to adjectives like awake

which can only qualify a noun when linked to it bya verb

the cat is awake for example The flex name and descrip

tion of this column are as follows

A For adjectives predicative Pred

ALemma Pred

The fourth column answers the question can this adjectival

lemma ever b e used in a p ostp ositiveway Here postpos

itive refers to lemmas like the adjective everlasting which

o ccurs in the phrase life everlasting here it comes after the

noun it mo dies whereas it is more normal in English for

a mo dier to come before the noun it qualies everlasting

life is also acceptable So everlasting gets the co de Y in

this column while adjectives like durable which never o ccur

p ostp ositivelygetthecode N

A For adjectives postpositive PostPos

PostPos ALemma

english linguistic guide

The fth and last column answers the question is this adjec

tive lemma only ever used in combination with certain other

words to make up a particular phrase An example is the

adjective batedwhich only ever o ccurs in the phrase with

bated breath and the adjective put whichonlyever o ccurs in

the phrase stayputSuch adjectives get the co de Y while all

other words get the co de NTheflex name and description

of this column are as follows

Exp A For adjectives expression

Exp ALemma

SUBCLASSIFICATION ADVERBS

There are ve adverb sub classication columns available

and they all take the form of Y N answers to questions as

describ ed in section ab ove They are presentedtoyou

in this add columns window

ADD COLUMNS

Ordinary

Predicative

Postpositive

Combinatory adverb

Expression

TOP MENU

PREVIOUS MENU

The rst of these ve columns answers the question is this

lemma an ordinary adverb where ordinary simply means

that it do esnt necessarily haveany sp ecial sub classication

features its just an adverb which can mo dify a verb or an

adjective Examples include generously which o ccurs in a

phrase like p eople havegiven very generously or a generously

illustrated b o ok The flex name and description of this

column are as follows

Ord ADV For adverbs ordinary

Ord ADVLemma

Sub classic atio n adverbs

The second of the veadverb sub classication column an

swers the question is this lemma an adverb which can usually

only b e used predicatively Adverbs which get the co de Y

here can b e distinguished from ordinary adverbs on the basis

of their predicative use with the verb be it is p ossible to say

the b oat is adrift but not the b oat is quicklyThus adrift

is a predicative adverb while quickly is not Typically many

predicativeadverbs b egin with the letter a around astern

or awry for example Manydonthowever high inland and

downtown amongst others All predicativeadverbs get the

co de Y in this column and all other lemmas get the co de N

The flex name and description of this column are as follows

Pred ADV For adverbs predicative

ADVLemma Pred

The third adverb sub classication column answers the ques

tion is this lemma an adverb that can be used postpos

itively Here p ostp ositive means after a noun as with

oside in the phrase several yards oside or apart in the

phrase a race apartAdverbs which can b e used in this way

get the co de Y in this column all other adverbs get the co de

N The flex name and description for this column are as

follows

PostPos ADV For adverbs postpositive

PostPos ADVLemma

The fourth of the ve adverb sub classication columns an

swers the question is this lemma an adverb whichcanbe

used in combination with a prep osition or another adverb

An example is the adverb cleanyou can say the sledgeham

mer broke clean through the do or here an adverb combines

with a prep osition Another example is the adverb allyou

can say they left the dog all alone here an adverb combines

with another adverb Adverbs which can combine in one of

these twoways get the co de Y all other lemmas get the co de

N The flex column name and description of this column are

as follows

For adverbs combinatory adverb ADV Comb

ADVLemma Comb

english linguistic guide

The last of the ve adjective sub classication columns an

swers the question is this adverb lemma only ever used in

combination with certain other words to make up a particular

phrase An example is the adverb amok whichonlyever

o ccurs in the phrase run amok and the adverb screamingly

which only ever o ccurs in the phrase screamingly funny

Suchadverbs get the co de Y while all other words get the

co de N The flex name and description of this column are

as follows

Exp ADV For adverbs expression

Exp ADVLemma

SUBCLASSIFICATION NUMERALS

The numerals given in the database are sucient to let you

sp ell anynumb er in full its p ossible to reconstruct the

orthographyofnumb er for example using nu

merals from the database As well as the numb ers them

selves a few extra terms suchasscore or umpteenth are

also given There are two sub classication columns available

for the numerals sp ecied in the database and they b oth

take the form of Y N answers to questions as explained in

section In addition there is a column which identies

those numerals used in certain expressions All three columns

are presented to you in this add columns window

ADD COLUMNS

Cardinal

Ordinal

Expression

TOP MENU

PREVIOUS MENU

The rst numeral sub classication column answers the ques

tion is this lemma a cardinal numb er Cardinal numbers

like three or seven or twelveare the most imp ortant forms

of numb ers and they simply indicate quantity rather than

rank order Any lemma in the database which is a cardinal

Sub classica tion numerals

numb er gets the co de Y in this column and every other

lemma gets the co de NTheflex name and description of

this column are as follows

For numerals cardinal Card NUM

NUMLemma Card

The second numeral sub classication column answers the

question is this lemma an ordinal number In contrast to

cardinal numb ers ordinalslike third or seventh or twelfth

indicate quantity and rank order Any lemma in the database

which is an ordinal numb er gets the co de Y in this column

and every other lemma gets the co de NThe flex name and

description of this column are as follows

Ord NUM For numerals ordinal

Ord NUMLemma

The last numeral sub classication column answers the ques

tion is this numeral ever used in combination with certain

other words to make up a particular phrase An example

is ninetyninewhich o ccurs in the phrase ninetynine times

out of a hundredandsixtyfour which o ccurs in the phrase

sixtyfour thousand dollar questionSuchnumerals get the

co de Y while all other words get the co de N The flex name

and description of this column are as follows

NUM For numerals expression Exp

NUMLemma Exp

SUBCLASSIFICATION PRONOUNS

There are a total of eight pronoun sub classication columns

available and they all take the form of Y N answers to

questions as explained in section They are presented

to you in these add columns windows

english linguistic guide

ADD COLUMNS

Personal

Demonstrative

Possessive

Reflexive

Whpronoun

Determinative use

TOP MENU

PREVIOUS MENU

v

ADD COLUMNS

Pronominal use

Expression

TOP MENU

PREVIOUS MENU

The rst of the seven pronoun sub classication columns an

swers the question is this lemma a p ersonal pronoun Pro

nouns which refer directly to p eople or things are p ersonal

pronouns This can include sub ject pronouns suchasthe

masculine third p erson singular form he and ob ject pro

nouns such as the third p erson plural form them These

lemmas thus get the co de Y all other lemmas get the co de

NTheflex name and description of this column are as

follows

Pers PRON For pronouns personal

Pers PRONLemma

The second of these columns answers the question is this

lemma a demonstrative pronoun The lemmas this that

these and those as used in phrases like that dress or this

sceptrd islegetthecode Y under this column all other

Sub classica tion pronouns

lemmas get the co de NTheflex name and description of

this column are as follows

PRON Dem For pronouns demonstrative

Dem PRONLemma

The third column answers the question is this lemma a

p ossessive pronoun Lemmas like her or hersasinher de

cision is hers and hers alone for example which can indicate

ownership of some sort b oth get the co de Y while all other

lemmas get the co de NTheflex name and description of

this column are as follows

Poss PRON For pronouns possessive

Poss PRONLemma

The fourth pronoun sub classication column answers the

question is this lemma a reexive pronoun Lemmas like

yourself in giveyourself a holiday or ourselves in wesawitfor

ourselves get the co de Y in this column all other lemmas get

the co de NTheflex name and description for this column

are as follows

Re PRON For pronouns reflexive

PRONLemma Re

The fth pronoun sub classication column answers the ques

tion is this lemma a whpronoun Such pronouns mostly

b egin with the letters wh and can b e used as relative pro

nouns an exp ert is one who knows more and more ab out less

and less orasinterrogative pronouns whodoyou love

All whpronouns suchaswho whither whence whoso ever

that howso ever and so on get the co de Y under this col

umn all other lemmas get the co de NTheflex name and

description of this column are as follows

PRON For pronouns whpronoun Wh

Wh PRONLemma

The sixth pronoun sub classication column answers the ques

tion can this pronoun lemma b e used as a determiner A

determiner helps to clarify which particular noun or noun

phrase you are referring to For example you might talk

english linguistic guide

ab out their cat in order to dierentiate it from your cat

or another cat As the name suggests suchwords help

determine what youre talking ab out Any lemma that can

b e used as a determiner gets the co de Y in this column and

every other lemma gets the co de NTheflex name and

description of this column are as follows

PRON Det For pronouns determinative use

Det PRONLemma

The seventh pronoun sub classication column answers the

question can this pronoun lemma b e used pronominally

where pronominally means instead of a noun or noun phrase

and indep endentofany other noun For example other can

b e used in a phrase like cho ose the otherormine in a phrase

like you cant its mine or mine is b etter so b oth these

pronouns can b e used pronominallyIncontrast my cannot

b e used pronominally it can only o ccur in phrases like my

book All pronoun lemmas which can b e used pronominally

get the co de Y in this column and all other lemmas get the

co de N The flex name and description of this column are

as follows

Pron PRON For pronouns pronominal use

PRONLemma Pron

The eighth and last pronoun sub classication columns an

swers the question is this pronoun lemma only ever used in

combination with certain other words to make up a particular

phrase Examples of this are few but one is aught which

o ccurs in the phrase for aughtIknow or for aught I care

Such pronouns get the co de Y while all other words get the

co de N The flex name and description of this column are

as follows

Exp PRON For adverbs expression

Exp PRONLemma

SUBCLASSIFICATION CONJUNCTIONS

There are two conjunction sub classication columns avail

able and they b oth take the form of Y N answerstoques

tions as describ ed in section ab ove They are presented

to you in this add columns window

Sub classica tion conjunctions

ADD COLUMNS

Coordinating

Subordinating

TOP MENU

PREVIOUS MENU

The rst column answers the question is this lemma a co

ordinating conjunction A conjunction like and is a co

ordinating conjunction it can link two or more clauses

together in suchaway that they remain of equal value to each

other or to put it another way one clause of the new sen

tence is not sub ordinate the other The other coordinating

conjunctions are but and or and these three lemmas get the

co de Y under this column All other lemmas get the co de N

The flex name and description of this column are as follows

C For conjunctions coordinating Cor

CLemma Cor

The second of the two conjunction sub classication columns

answers the question is this conjunction a sub ordinating

conjunction A conjunction like because is a sub ordinating

conjunction it can link two clauses together in suchaway

that one clause b ecomes dep endent on the other as in a

sentence like hes reading a b o ok b ecause the television is

broken Here because acts as a sub ordinating conjunction

the link b etween the two clauses is more complex than the

use of a coordinating conjunction like and would imply All

conjunction lemmas which are sub ordinating conjunctions

get the co de Y in this column all other lemmas get the

co de N The flex name and description of this column are

as follows

Sub C For conjunctions subordinating

Sub CLemma

english linguistic guide

ENGLISH FREQUENCY

The frequency information given in the database that is

details of how often words o ccur in English is available b oth

for lemmas and wordforms It is taken from the cobuild

corpus of the University of Birmingham which in the early

version extracted for and corrected by celex contained

ab out million words taken from written sources of many

kinds and some sp oken sources as well Frequency gures

are available for lemmas and for wordforms

The starting p oint for calculating frequency information is

the cobuild million word corpus a countismadeof

the numb er of times each string o ccurs This task is easy

for a computer which can quickly makea count of all the

words that app ear in the corpus The resulting gures are

raw string counts that is they indicate how many times

each separate group of letters o ccurs in the corpus taking no

account of the dierent meanings or word classes that can

b e applied to eachgroup You can see the remaining raw

counts in a cobuild corpus typ es lexicon when you select

the Freq column The string countoffamilies for example

is while for bank it is Todevelop this basic

string countinto a more helpful word count the strings must

be identied either as wordforms which can b e linked to a

particular lemma or as other things not represented in the

database such as p ersonal names foreign words and words

mistyp ed or misread by an Optical Character Recognition

machine

Sometimes this identication pro cess is straightforward the

string millstones is only ever the plural wordform of the noun

lemma millstone So in this case the raw string frequency of

the string millstones is also the frequency of the wordform

millstones and so in the wordform lexicon Cob column it

gets the same frequency as the string

Once you know the frequencies of the wordforms asso ciated

with a particular lemma working out a frequency gure for

the lemma as a whole is straightforward all you havetodo

is add up the appropriate wordform frequencies In this way

English frequency

the frequency of the noun lemma millstone is the frequency of

the wordform millstones plus the frequency of the wordform

millstone The frequency of the lemma millstone is the total

of the two and this is the gure given in the lemma lexicon

Cob column

The only way to sort out the individual frequencies of eachof

these strings is to lo ok at the way they are used in the corpus

a pro cess known as disambiguation Its p ossible to carry

out this task quickly by computer program but at present

the results of such programs can never b e wholly accurate

For this reason celex chose to disambiguate by hand which

means that someone reads each o ccurrence of eachambiguous

form in the corpus and notes the lemma to which it b elongs

While such an approach is b oth costly and timeconsuming

it do es pro duce results which are more dep endable and accu

rate For jump er it seems that of the o ccurrences mean

item of clothing and mean someone who jumps These are

the two gures given in the wordform lexicon Cob column

for the twodierent jump er wordforms Sometimes not all

o ccurrences refer to wordforms in the database Some maybe

prop er nouns surnames for example or typing errors and

some simply cant b e disambiguated For example jump er

o ccurs times in relation to a p ersons name Such infor

mation is not given in the database since it do esnt relate

directly to any of the lemmas or wordforms available

Again once the wordform frequencies have b een claried

working out the lemma frequencies is straightforward For

the two lemmas with the form jump er the lemma frequencies

are meaning clothing and meaning someone who

jumps giving a total of These lemma frequency gures

are given in the lemma lexicon Cob column and in the same

column to b e found with the lemma information given for

wordforms

When strings o ccur very frequently in the corpus the work

required to disambiguate eachcasebyhandcanbedaunting

It may also b e unnecessary since an intelligent estimate cou

pled with an indication of how far that estimate is accurate

should usually b e enough So whenever ambiguous words o c

cur more than times in the corpus not all the o ccurrences

are checked individually Instead one hundred o ccurrences

of the string are taken at random from the corpus and then

analysed In this way its p ossible to formulate a ratio which

english linguistic guide

indicates the prop ortions of the various interpretations and

this ratio can then b e applied to the real gures to see an

estimate of how the fully disambiguated gures would lo ok

As an example take the string bank Its basic corpus string

frequency is It can either b e a singular noun or an

instance of a verb the rst or second p erson singular form

the plural form or the innitive Here is a lexicon which

shows these wordforms with their word class and frequency

Word Class Cob

bank N

bank V

bank V

bank V

bank V

To calculate these gures a o ccurrences of the string

bank were taken from the corpus and disambiguated by hand

It turned out that of the o ccurrences b elonged to the verbal

lemma and to the noun lemma So to estimate the real

frequency of the wordform b elonging to the noun lemma

divide the numb er of times it o ccurred in the sample by

the total numb er of successfully disambiguated forms and

then multiply the result by the original string frequency

Rep eating this pro cedure gives for the

verb lemma This latter gure is divided equally for the

four p ossible wordforms each for the rst p erson singular

form the second p erson singular form the plural form and

the innitive This is the usual way of sorting out ambiguous

verbal ections since disambiguating every verbal form by

hand is a task whichwould involve a great deal of work

yielding results of interest to only a few

For most items in the database the frequency gures are

accurate However when estimates have to b e made on the

basis of a hundred examples then deviation gures haveto

b e calculated to let you see just how accurate the estimates

are This formula gives the required deviation gure

r

p
p N
n

N

n N

where N is the frequency of the string as a whole n is

the numb er of items which could b e disambiguated in the

English frequency

random item sample and p is the ratio gure for the

item when it b elongs to one particular lemma Thus for the

noun wordform bank N is n is and p is The

formula gives as the deviation This means that the true

frequency for this form of bank is almost certainat least

certainto lie b etween and

Word ClassLemma Cob CobDev

bank N

bank V

bank V

bank V

bank V

Whenever the deviation is greater than the frequency itself

then you know for sure that some sort of arbitrary approx

imation has b een carried out This happ ened for the verbal

forms of bankasyou can see in the table ab ove

Working out deviation gures for a lemma involves adding to

gether the frequencies of its disambiguated wordforms And

once again whenever the resulting deviation gure is equal

to or greater than the frequency itself you know that some

arbitrary disambiguation has b een necessary

One nal p oint to note here is that some frequency infor

mation is available with the orthographic columns This

relates directly to the dierent sp ellings that wordforms or

headwords can have It do es not aect the frequency infor

mation given here which deals with each form as a whole

regardless of how it can b e sp elt For instance realize can

also b e sp elt realise and the lemma frequency given alongside

each of them is the same The sp elling frequency

on the other hand shows that the sp elling realize o ccurs

times while the sp elling realise o ccurs For more

details ab out this extra layer of disambiguation read the

appropriate subsection under English Orthography

FREQUENCY INFORMATION FOR LEMMAS AND

WORDFORMS

Now that the background details have b een explained the

individual column names and descriptions can b e formally

dened For b oth lemmas and wordforms there are four

english linguistic guide

columns available which express the cobuild frequency g

ures in various ways

The rst column gives the plain cobuild frequency count

for each lemma or wordform The gure given in the lemma

version of the column for collate is which means that out

of the words in the corpus are the word collate

in some form or other The gures given in the wordform

version of this column reveal how frequently eachofthe

p ossible forms o ccur for collate the gure is for collates

it is There are also o ccurrences of collatedand

o ccurrences of collating The flex name and description of

this column are as follows

Cob COBUILD frequency

CobLemma

The second column indicates how accurate the frequencies

in the previous column are byproviding a deviation gure

for each lemma or wordform calculated according to the

metho ds describ ed in the previous section If a word has

b een fully disambiguated without the need for any estimates

the gure is When some estimation has b een required the

gure will b e greater than zero If the gure should ever b e

equal to or greater than the frequency it qualies then you

know that full disambiguation was not p ossible The gure

given for the verb lemma shine in the sense of b e bright or

direct the light is and when you use it in conjunction

with the cobuild frequency gure of it indicates that

you can b e almost certain certain that shine o ccurs in

one form or another somewhere b etween and times

The flex name and description of this column are as follows

CobDev COBUILD frequency deviation

CobDevLemma

The next column contains the same frequency gures as the

rst column except that they havebeenscaleddown to a

range of to instead of the usual to

This is done by dividing the normal cobuild frequency for

eachword bythenumber of words in the whole corpus and

then multiplying the answer by The end result

is a set of gures which are probably easier to understand

it makes greater sense to say that the word magnicent is

Frequency information for lemmas and wordforms

twenty in a million than it do es to say that its words

out of And since other wellknown text corp ora

suchastheLondonOsloBergen lob and Brown corp ora of

Englishare also based on a count of one million this scale

provides the opp ortunityforinteresting comparisons to b e

made However as you might exp ect some detail is lost in the

scalingdown pro cess the words barb ecue and babysitter

whichhave the million word lemma frequencies of and

resp ectively b oth share the same million word frequency

of

CobMln COBUILD frequency

CobMlnLemma

For those whose work requires a further transformation of the

gures psycholinguists working with stimulus resp onse times

for example a column containing logarithmic values is avail

able The eect of the logarithmic scale is to emphasize the

imp ortance of lower frequency words in a way that the usual

linear scale do es not For example the dierence b etween

twowords one of frequency and the other of frequency

b ecomes much greater than the dierence b etween twowords

of frequency and For the rst pair of words the

dierence is while for the second pair the dierence

is a mere This conrms mathematicallywhatwe

knowintuitively b ecause there are so manywords with a

low frequency the dierences b etween them are that much

more signicant With a high frequency word a dierence

of one or two isnt very signicant

The values given are the base logarithms of each COBUILD

frequency describ ed ab ove In place of a

scale from to the resulting logarithmic values

in this column range from zero log to log

And when a word has a normal frequency of zero the log

arithmic value is also given as zero This is mathemati

cally inaccurate log do esnt exist butat least in this

x

contextrelatively unimp ortant anyword with a logarith

mic frequency of o ccurs at the very most only times

in the full cobuild million word corpus The thing to

rememb er is that only words whichhavea cobuild

frequency value of twoormoreorifyou prefer only words

which o ccur or more times in the cobuild corpus have

a logarithmic value greater than zero

english linguistic guide

CobLog COBUILD frequency logarithmic

CobLogLemma

FREQUENCY INFORMATION FROM WRITTEN AND

SPOKEN SOURCES

Ab out words in the cobuild corpus makeup

written texts and the remaining words makeup

sp oken texts In a sense then there are two other corp ora

you can use one which deals with written texts only and one

with sp oken texts onlyYou can cho ose for yourself whether

you wish to use either written or sp oken gures in place of

the full gures explained in the preceeding sections The

metho ds used in working out the gures given are the same

as those describ ed in the previous section

The columns available for written and sp oken corpus fre

quencies are roughly the same as those for the full corpus

with the exception of the deviation gures they are not

recalculated for the written and sp oken texts Instead you

can use the gures given for the full corpus though remember

that when you apply them to frequencies for the written and

sp oken corp ora the range of error is actually larger than

would otherwise b e

WRITTEN CORPUS INFORMATION

There are three columns whichcontain frequency information

for the written sources in the cobuild corpus The gure

given in the lemma version of the column for memory is

which means that out of the words in the

corpus are the word memory in some form or other

The gures given in the wordform version of this column

reveal how frequently each of the p ossible forms o ccur for

memory the gure is and for memories it is The

flex name and description of this column are as follows

CobW COBUILD written frequency m

CobWLemma

The next column contains the same frequency gures as

CobW except that they havebeenscaleddown to a range of to instead of the usual to This

Written corpus information

is done by dividing the normal cobuild written frequency

for eachword by the number of words in the written cor

pus ab out and then multiplying the answer

by The end result is a set of gures which are

probably easier to understand it makes greater sense to say

that a word is one in a million than it do es to say that its

words out of However as you might exp ect some

detail is lost in the scalingdown pro cess all words which

have million word lemma frequencies of b etween and

share the same million word frequency of

CobWMln COBUILD written frequency

CobWMlnLemma

The third and last written corpus column contains the base

logarithms of each CobWMln for the reasons describ ed

ab ove in connection with the full corpus In place of a scale

from to then the resulting logarithmic values

in this column range from zero log to log

And when a word has a normal frequency of zero the log

arithmic value is also given as zero This is mathemati

cally inaccurate log do esnt exist butat least in this

x

contextrelatively unimp ortant anyword with a logarith

mic frequency of o ccurs at the verymostonlytimesin

the cobuild million written word corpus The thing

to remember is that only words whichhavea CobWMln

frequency value of twoormoreorifyou prefer only words

which o ccur or more times in the cobuild corpus have

a logarithmic value greater than zero

CobWLog COBUILD written frequency logarithmic

CobWLogLemma

SPOKEN CORPUS INFORMATION

There are three columns whichcontain frequency information

for the sp oken sources in the cobuild corpus The gure

given in the lemma version of the column for memory is

which means that out of the approximately words

in the corpus are the word memory in some form or other

The gures given in the wordform version of this column

reveal how frequently each of the p ossible forms o ccur for

english linguistic guide

memory the gure is and for memories it is The flex

name and description of this column are as follows

CobS COBUILD spoken frequency m

CobSLemma

The next column contains the same frequency gures as

CobS except that they have b een scaled down to a range

of to instead of the usual to This is

done by dividing the normal cobuild sp oken frequency for

eachword by the number of words in the sp oken corpus and

then multiplying the answer by

CobSMln COBUILD spoken frequency

CobSMlnLemma

The third and last sp oken corpus column contains the base

logarithms of each CobSMln frequency for the reasons

describ ed ab ove in connection with the full corpus In place

of a scale from to the resulting logarithmic values

in this column range from zero log to log

Andwhenaword has a normal frequency of zero the loga

rithmic value is also given as zero This is mathematically in

accurate log do esnt exist butat least in this context

x

relatively unimp ortant anyword with a logarithmic fre

quency of o ccurs at the very most only once in the cobuild

million sp oken word corpus and is consequently only of

interest to those concerned with the more esoteric branches

of lexicography The thing to rememb er is that only words

whichhaveanCobSMln frequency value of two or more or

if you prefer only words which o ccur two or more times in

the cobuild sp oken corpus have a logarithmic value greater

than zero

CobSLog COBUILD spoken frequency logarithmic

CobSLogLemma

FREQUENCY INFORMATION FOR COBUILD

CORPUS TYPES

The frequency information given in cobuild corpus typ es

lexicons consists of the raw string counts from which all

the other frequency gures for lemmas wordforms and ab

breviations are derived Also available are gures for the

Frequency information for COBUILD corpus typ es

sp oken and written texts in the corpus as well as for British

English and American English typ es whicharenottobe

found amongst the wordforms and abbreviations given in the

celex database If you are not already familiar with the

terms token and typ ethencheck the glossary and the rst

part of the manual the Intro duction in the section Lexicon

typ es

The rst column simply lists the orthographic forms of all

typ es as they o ccur in the cobuild corpus The flex name

and description of this column are as follows

Typ e Graphemic transcription

The second column is the basic string count which tells you

how many times eachtyp e o ccurs in the cobuild corpus

whichcontains ab out tokens The flex name

and description of this column are as follows

Freq Absolute frequency

FREQUENCY INFORMATION FOR COBUILD

WRITTEN CORPUS TYPES

There are four columns which contain raw string counts from

the written texts in the cobuild corpus The rst contains

the frequencies of all typ es which o ccur more than once in

all the written texts

FreqW COBUILD written frequency m

The next column contains raw string counts from the written

texts that are normal British usages as opp osed to American

English or some other brand of English

FreqWB COBUILD written frequency British English

The third column contains raw string counts from the written

texts that are normal American usages as opp osed to British

English or some other brand of English

FreqWA COBUILD written frequency American English

english linguistic guide

The fourth column contains raw string counts from the writ

ten texts that are not normal British or American usages

but are instead from an unidentied brand of English

FreqWU COBUILD written frequency undetermined origin

FREQUENCY INFORMATION FOR COBUILD

SPOKEN CORPUS TYPES

There are three columns whichcontain raw string counts

from the sp oken texts in the cobuild corpus Ab out

million words were transcrib ed from recorded conversations

and included in the corpus None of conversations tran

scrib ed involved American English so no separate gures

for American English sp oken typ es are available

The rst column contains the frequencies of all typ es which

o ccur more than once in the sp oken texts

FreqS COBUILD spoken frequency m

The next column contains raw string counts from the written

texts that are marked as normal British usages

FreqSB COBUILD spoken frequency British English

The third column contains raw string counts from the written

texts that are not normal British usages but are instead from

an unidentied brand of English

FreqSU COBUILD spoken frequency undetermined origin

TREE DIAGRAMS COLUMN

DESCRIPTIONS

This app endix is divided into sections corresp onding to the

lexicon typ es currently availableEach section b egins with a

set of tree diagrams whichgiveyou an overview of the columns

you can cho ose when you select a particular typ e of lexicon and

then there are technical details ab out each of those columns

thetyp e of the column its minimum and maximum values

and lengths the number of null values it contains and the

characters used in each column These details are particularly

useful when you exp ort a le from flex

Whenever a new version of the database is released the cor

resp onding section in this app endix will also b e replaced with

the relevant diagrams and technical details Always remember

to check the name and lexicon numb er when youre using this

app endix you can see which lexicon typ e and version you are

dealing with by reading the title of each diagram or the line at

the top of each righthand page

ORTHOGRAPHY OF ENGLISH LEMMAS E

Numb er of sp ellings OrthoCnt

Sp elling numb er N OrthoNum

Language co de OrthoStatus

COBUILD frequency m CobSp ellFreq

Frequency of sp elling

COBUILD condence deviation m CobSp ellDev

Without diacritics Head

Without diacritics reversed HeadRev

With diacritics HeadDia

Plain

Purely lowercase alphab etical HeadLow

Purely lowercase alphab etical sorted HeadLowSort

Sp elling Numb er of letters HeadCnt

Without diacritics HeadSyl

Syllabied With diacritics HeadSylDia

Numb er of syllables HeadSylCnt

PHONOLOGY OF ENGLISH LEMMAS E

Numb er of pronunciations PronCnt

PronNum Pronunciation numb er N

Status of pronunciation PronStatus

SAMPAchar set PhonSAM

CELEX char set PhonCLX

Plain CPAchar set PhonCPA

DISC char set PhonDISC

Numb er of phonemes PhonCnt

SAMPAchar set PhonSylSAM

CELEX char set PhonSylCLX

CELEX char set brackets PhonSylBCLX

Phonetic transcriptions Syllabied

CPAchar set PhonSylCPA

DISC char set PhonSylDISC

Numb er of syllables SylCnt

SAMPAchar set PhonStrsSAM

CELEX char set PhonStrsCLX

Syllabied

with stress CPAchar set PhonStrsCPA

DISC char set PhonStrsDISC

Stress pattern StrsPat

CV pattern PhonCV

Phonetic patterns Syllabied

CV pattern brackets PhonCVBr

MORPHOLOGY OF ENGLISH LEMMAS E

Status MorphStatus

Lang Language

information

Numb er of morphological analyses MorphCnt

Analysis numb er N MorphNum

Nounverbax comp ound NVAComp

Derivation metho d Der

Status of morphological analyses Comp ound metho d Comp

Deriv comp ound metho d DerComp

Default analysis Def

Stems axes Imm

Class lab els ImmClass

Class verb sub cat lab els ImmSubCat

Stemax lab els ImmSA

Stem allomorphy ImmAllo

Immediate

segmentation Ax substitution ImmSubst

Opacity ImmOpac

Derivational transformation TransDer

Inxation ImmInx

Reversion ImmRevers

Stems axes Flat

Complete Derivational

comp ositional Segmentations segmentation Class lab els FlatClass

information at

Stemax lab els FlatSA

Stems axes Struc

Stems axes lab elled StrucLab

Complete Empty brackets lab elled StrucBrackLab

segmentation

hierarchical Stem allomorphy StrucAllo

Ax substitution StrucSubst

Opacity StrucOpac

Numb er of comp onents CompCnt

Other Numb er of morphemes MorCnt

Number of levels LevelCnt

SYNTAX OF ENGLISH LEMMAS

NOUNS VERBS E

Numeric co des ClassNum

Word class

Lab els Class

Count C N

Uncount Unc N

Singular use Sing N

Plural use Plu N

Group count GrC N

Sub classication Group uncount GrUnC N

nouns

Attributive Attr N

Postp ositive PostPos N

Vo cative Voc N

Prop er noun Prop er N

Expression Exp N

Transitive Trans V

Transitive complementation TransComp V

Intransitive Intrans V

Ditransitive Ditrans V

Sub classication Linking verb Link V

verbs

Phrasal Phr V

Prep ositional Prep V

Phrasal prep ositional PhrPrep V

Expression Exp V

SYNTAX OF ENGLISH LEMMAS

ADJECTIVES ADVERBS NUMERALS

PRONOUNS CONJUNCTIONS E

Ordinary Ord A

Attributive Attr A

Predicative Pred A

Sub classication

adjectives Postp ositive PostPos A

Group uncount GrUnc A

Expression Exp A

Ordinary Ord ADV

Predicative Pred ADV

Sub classication Postp ositive PostPos ADV

adverbs

Combinatory Comb ADV

Expression Exp ADV

Cardinal Card NUM

Sub classication Ordinal Ord NUM

numerals

Expression Exp NUM

Personal Pers PRON

Demonstrative Dem PRON

Possessive Poss PRON

Reexive Re PRON

Sub classication

pronouns Whpronoun Wh PRON

Determinativeuse Det PRON

Pronominal use Pron PRON

Expression Exp PRON

Co ordinating Cor C

Sub classication

conjunctions Sub ordinating Sub C

FREQUENCY OF ENGLISH LEMMAS E

COBUILD frequency m Cob

COBUILD condence deviation m CobDev

COBUILD all sources

COBUILD frequency m CobMln

COBUILD frequency logarithmic CobLog

COBUILD written frequency m CobW

COBUILD written sources COBUILD written frequency m CobWMln

COBUILD written frequency logarithmic CobWLog

COBUILD sp oken frequency m CobS

COBUILD sp oken sources COBUILD sp oken frequency m CobSMln

COBUILD sp oken frequency logarithmic CobSLog

Appendix

Attr A For adjectives attributive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Attr N For nouns attributive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

C N For nouns countable

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Card NUM For numerals cardinal

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Class Word class labels

Typ e character Null values

Minimum value A Minimum length

Maximum value V Maximum length

Characters ACDEIM NOPRSTU V

Column descriptions for English Lemmas E

ClassNum Word class numeric

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Cob COBUILD frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobDev COBUILD frequency deviation

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobLog COBUILD frequency logarithmic

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobMln COBUILD frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Appendix

CobS COBUILD spoken frequency m

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobSLog COBUILD spoken frequency logarithmic

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobSMln COBUILD spoken frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobSp ellDev COBUILD spelling frequency deviation

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobSp ellFreq COBUILD spelling frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Column descriptions for English Lemmas E

CobW COBUILD written frequency m

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobWLog COBUILD written frequency logarithmic

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobWMln COBUILD written frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Comb ADV For adverbs combinatory

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Comp Compound analysis

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Appendix

CompCnt Number of morphological components

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Cor C For conjunctions coordinating

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Dem PRON For pronouns demonstrative

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Der Derivation analysis

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

DerComp Derivational compound analysis

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Column descriptions for English Lemmas E

Det PRON For pronouns determinative use

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Ditrans V For verbs ditransitive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Exp A For adjectives expression

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Exp ADV For adverbs expression

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Exp N For nouns expression

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Appendix

Exp NUM For numerals expression

Typ e character Null values

Minimum value N Minimum length

Maximum value N Maximum length

Characters N

Exp PRON For pronouns expression

Typ e character Null values

Minimum value N Minimum length

Maximum value N Maximum length

Characters N

Exp V For verbs expression

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Flat Flat segmentation

Typ e character Null values

Minimum value April Minimum length

Maximum value zoom Maximum length

Characters ABD EFGHIJL MOPQSTUV Wa

bcdefg hijklmn opqrstuv wx yz

Column descriptions for English Lemmas E

FlatClass Flat segmentation word class labels

Typ e character Null values

Minimum value A Minimum length

Maximum value xxxx Maximum length

Characters ABCDIN OPQSTVx

FlatSA Flat segmentation stemaffix labels

Typ e character Null values

Minimum value AA Minimum length

Maximum value SSSA Maximum length

Characters AFS

GrC N For nouns group countable

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

GrUnc N For nouns group uncountable

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Appendix

Head Headword

Typ e character Null values

Minimum value d Minimum length

Maximum value zoophyte Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs

tuvwxy z

HeadCnt Headword number of letters

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

HeadDia Headword diacritics

Typ e character Null values

Minimum value d Minimum length

Maximum value epee Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs

tuvwxy zaaceeenou

HeadLow Headword lowercase alphabetical

Typ e character Null values

Minimum value a Minimum length

Maximum value zoophyte Maximum length

Characters abcdef ghijklm nopqrstu vw xyz

Column descriptions for English Lemmas E

HeadLowSort Headword lowercase alphabetical sorted

Typ e character Null values

Minimum value a Minimum length

Maximum value ttuu Maximum length

Characters abcdef ghijklm nopqrstu vw

xyz

HeadRev Headword reversed

Typ e character Null values

Minimum value oht Minimum length

Maximum value zzuf Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs

tuvwxy z

HeadSyl Headword syllabified

Typ e character Null values

Minimum value d Minimum length

Maximum value zoom Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs

tuvwxy z

HeadSylCnt Headword number of orthographic syllables

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Appendix

HeadSylDia Headword syllabified diacritics

Typ e character Null values

Minimum value d Minimum length

Maximum value epee Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs

tuvwxy zaaceeenou

IdNum Lemma number

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Imm Immediate segmentation

Typ e character Null values

Minimum value April Minimum length

Maximum value zoom Maximum length

Characters ABD EFGHIJL MOPQSTUV Wa

bcdefg hijklmn opqrstuv wx

yz

ImmAllo Stem allomorphy top level

Typ e character Null values

Minimum value B Minimum length

Maximum value Z Maximum length

Characters BCDFNZ

Column descriptions for English Lemmas E

ImmClass Immediate segmentation word class labels

Typ e character Null values

Minimum value A Minimum length

Maximum value xxx Maximum length

Characters ABCDIN OPQSTVx

ImmInx Infixation top level

Typ e character Null values

Minimum value N Minimum length

Maximum value N Maximum length

Characters N

ImmOpac Opacity top level

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

ImmRevers Reversion top level

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

ImmSA Immediate segmentation stemaffix labels

Typ e character Null values

Minimum value AA Minimum length

Maximum value SSA Maximum length

Characters AFS

Appendix

ImmSubCat Immediate segmentation subcat labels

Typ e character Null values

Minimum value Minimum length

Maximum value xxx Maximum length

Characters AB CDINOPQ STx

ImmSubst Affix substitution top level

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Intrans V For verbs intransitive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Lang Language information

Typ e character Null values

Minimum value A Minimum length

Maximum value S Maximum length

Characters ABDFGI LS

LevelCnt Number of morphological levels

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Column descriptions for English Lemmas E

Link V For verbs linking verb

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

MorCnt Number of morphemes

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

MorphCnt Number of morphological analyses

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

MorphNum Morphological analysis number

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

MorphStatus Morphological status

Typ e character Null values

Minimum value C Minimum length

Maximum value Z Maximum length

Characters CFIMOR UZ

Appendix

NVAComp Nounverbaffix compound

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Ord A For adjectives ordinary

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Ord ADV For adverbs ordinary

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Ord NUM For numerals ordinal

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

OrthoCnt Number of spellings

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Column descriptions for English Lemmas E

OrthoNum Spelling number

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

OrthoStatus Status of spelling

Typ e character Null values

Minimum value A Minimum length

Maximum value B Maximum length

Characters AB

Pers PRON For pronouns personal

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

PhonCLX Phon headword CELEX charset

Typ e character Null values

Minimum value NSs Minimum length

Maximum value zum Maximum length

Characters ADEINO STUVZabd ef ghijkl mnprstu vwxz

Appendix

PhonCnt Headword number of phonemes

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

PhonCPA Phon headword CPA charset

Typ e character Null values

Minimum value Minimum length

Maximum value zum Maximum length

Characters ADEIJNO STUZabd ef

ghijkl mnoprst uvwxz

PhonCV Headword phon CV pattern

Typ e character Null values

Minimum value C Minimum length

Maximum value VVCCCVCC Maximum length

Characters CSV

PhonCVBr Headword phon CV pattern with brackets

Typ e character Null values

Minimum value CCCVCCC Minimum length

Maximum value VVC Maximum length

Characters CSV

Column descriptions for English Lemmas E

PhonDISC Phon headword DISC charset

Typ e character Null values

Minimum value Minimum length

Maximum value tt Maximum length

Characters CDEFHIJN PQ

RSTUVZ bcdfgh ijklmnpq rs

tuvwxz

PhonSAM Phon headword SAMPA charset

Typ e character Null values

Minimum value T Minimum length

Maximum value nZeInju Maximum length

Characters ADEINOQ STUVZabd ef

ghijkl mnprstu vwxz

PhonStrsCLX Syll phon headword with stress CELEX charset

Typ e character Null values

Minimum value blISnIst Minimum length

Maximum value zUOldZIst Maximum length

Characters ADEI NOSTUVZa bd

efghij klmnprs tuvwxz

PhonStrsCPA Syll phon headword with stress CPA charset

Typ e character Null values

Minimum value mOrl Minimum length

Maximum value zUOlJIst Maximum length

Characters ADEIJ NOSTUZa bd efghij klmnopr stuvwxz

Appendix

PhonStrsDISC Syll phon headword with stress DISC charset

Typ e character Null values

Minimum value mEn Minimum length

Maximum value ntt Maximum length

Characters CDEFH IJ

NPQRST UVZbcd fghijklm np

qrstuv wxz

PhonStrsSAM Syll phon headword with stress SAMPA charset

Typ e character Null values

Minimum value TI Minimum length

Maximum value zbEstQs Maximum length

Characters ADEIN OQSTUVZa bd

efghij klmnprs tuvwxz

PhonSylCLX Syll phon headword CELEX charset

Typ e character Null values

Minimum value SI Minimum length

Maximum value zum Maximum length

Characters ADEINO STUVZabd ef

ghijkl mnprstu vwxz

PhonSylBCLX Syll phon headword CELEX charset brackets

Typ e character Null values

Minimum value NSs Minimum length

Maximum value zum Maximum length

Characters ADEINOS TUVZab de fghijk lmnprst uvwxz

Column descriptions for English Lemmas E

PhonSylCPA Syll phon headword CPA charset

Typ e character Null values

Minimum value Minimum length

Maximum value zum Maximum length

Characters ADEIJNO STUZabd ef

ghijkl mnoprst uvwxz

PhonSylDISC Syll phon headword DISC charset

Typ e character Null values

Minimum value Minimum length

Maximum value ntt Maximum length

Characters CDEFHIJ NP

QRSTUV Zbcdfg hijklmnp qr

stuvwx z

PhonSylSAM Syll phon headword SAMPA charset

Typ e character Null values

Minimum value TI Minimum length

Maximum value nZeInju Maximum length

Characters ADEINOQ STUVZabd ef

ghijkl mnprstu vwxz

Phr V For verbs phrasal verb

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Appendix

PhrPrep V For verbs phrasal prepositional verb

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Plu N For nouns plural use

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Poss PRON For pronouns possessive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

PostPos A For adjectives postpositive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

PostPos ADV For adverbs postpositive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Column descriptions for English Lemmas E

PostPos N For nouns postpositive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Pred A For adjectives predicative

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Pred ADV For adverbs predicative

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Def Default analysis

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Prep V For verb prepositional verb

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Appendix

Pron PRON For pronouns pronominal use

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

PronCnt Number of pronunciations

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

PronNum Pronunciation number

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

PronStatus Status of pronunciation

Typ e character Null values

Minimum value P Minimum length

Maximum value S Maximum length

Characters PS

Prop er N For nouns proper noun

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Column descriptions for English Lemmas E

Re PRON For pronouns reflexive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Sing N For nouns singular use

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

StrsPat Headword stress pattern

Typ e character Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Struc Structured segmentation

Typ e character Null values

Minimum value confideent Minimum length

encetrick

Maximum value zoom Maximum length

Characters A BDEFGHI JLMOPQST UV

Wabcde fghijkl mnopqrst uv wxyz

Appendix

StrucAllo Stem allomorphy any level

Typ e character Null values

Minimum value B Minimum length

Maximum value Z Maximum length

Characters BCDFNZ

StrucBrackLab Structured segmentation word class labels only

Typ e character Null values

Minimum value VN Minimum length

Maximum value V Maximum length

Characters A BCDINOP QSTVx

StrucLab Structured segmentation word class labels

Typ e character Null values

Minimum length Minimum value confideVent

AVAence

NAN

trickVNN

NV

Maximum value zoomV Maximum length

Characters ABCDEFG HIJLMNOP QS

TUVW abcdefg hijklmno pq

rstuvw xyz

StrucOpac Opacity any level

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Column descriptions for English Lemmas E

StrucSubst Affix substitution any level

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Sub C For conjunctions subordinating

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

SylCnt Headword number of phonetic syllables

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Trans V For verbs transitive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

TransComp V For verbs transitive plus complementation

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Appendix

TransDer Derivational transformation top level

Typ e character Null values

Minimum value Minimum length

Maximum value yuponi Maximum length

Characters abc defghik lmnopqrs tu

vwxyz

Unc N For nouns uncountable

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Voc N For nouns vocative

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Wh PRON For pronouns whpronoun

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

ORTHOGRAPHY OF ENGLISH WORDFORMS E

Numb er of sp ellings OrthoCnt

Sp elling numb er N OrthoNum

Language co de OrthoStatus

COBUILD frequency m CobSp ellFreq

Frequency of sp elling

COBUILD condence deviation m CobSp ellDev

Without diacritics Word

Without diacritics reversed WordRev

With diacritics WordDia

Plain

Purely lowercase alphab etical WordLow

Purely lowercase alphab etical sorted WordLowSort

Sp elling Numb er of letters WordCnt

Without diacritics WordSyl

Syllabied With diacritics WordSylDia

Numb er of syllables WordSylCnt

PHONOLOGY OF ENGLISH WORDFORMS E

Numb er of pronunciations PronCnt

PronNum Pronunciation numb er N

Status of pronunciation PronStatus

SAMPAchar set PhonSAM

CELEX char set PhonCLX

Plain CPAchar set PhonCPA

DISC char set PhonDISC

Numb er of phonemes PhonCnt

SAMPAchar set PhonSylSAM

CELEX char set PhonSylCLX

CELEX char set brackets PhonSylBCLX

Phonetic transcriptions Syllabied

CPAchar set PhonSylCPA

DISC char set PhonSylDISC

Numb er of syllables SylCnt

SAMPAchar set PhonStrsSAM

CELEX char set PhonStrsCLX

Syllabied

with stress CPAchar set PhonStrsCPA

DISC char set PhonStrsDISC

Stress pattern StrsPat

CV pattern PhonCV

Phonetic patterns Syllabied

CV pattern brackets PhonCVBr

MORPHOLOGY OF ENGLISH WORDFORMS E

Numeric id IDNumLemma

Orthography ORTHOGRAPHY OF ENGLISH LEMMAS

Phonology PHONOLOGY OF ENGLISH LEMMAS

Lemma

information Morphology MORPHOLOGY OF ENGLISH LEMMAS

Syntax SYNTAX OF ENGLISH LEMMAS

Frequency FREQUENCY OF ENGLISH LEMMAS

See the information

in these diagrams for

the available columns

Singular Sing

Plural Plu

Positive Pos

Comparative Comp

Sup erlative Sup

Innitive Inf

Inectional Participle Part

features

Presenttense Pres

Past tense Past

st p erson verb Sin

nd p erson verb Sin

rd p erson verb Sin

Rare form Rare

Typ e of ection FlectTyp e

TransIn Inectional

Transformation

FREQUENCY OF ENGLISH WORDFORMS E

COBUILD frequency m Cob

COBUILD condence deviation m CobDev

COBUILD all sources

COBUILD frequency m CobMln

COBUILD frequency logarithmic CobLog

COBUILD written frequency m CobW

COBUILD written sources COBUILD written frequency m CobWMln

COBUILD written frequency logarithmic CobWLog

COBUILD sp oken frequency m CobS

COBUILD sp oken sources COBUILD sp oken frequency m CobSMln

COBUILD sp oken frequency logarithmic CobSLog

Column descriptions for English wordforms E

Cob COBUILD frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobDev COBUILD frequency deviation

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobLog COBUILD frequency logarithmic

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobMln COBUILD frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobS COBUILD spoken frequency m

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Appendix

CobSLog COBUILD spoken frequency logarithmic

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobSMln COBUILD spoken frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobSp ellDev COBUILD spelling frequency deviation

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobSp ellFreq COBUILD spelling frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobW COBUILD written frequency m

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Column descriptions for English wordforms E

CobWLog COBUILD written frequency logarithmic

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

CobWMln COBUILD written frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Comp Inflectional feature comparative

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

FlectTyp e Type of flection

Typ e character Null values

Minimum value P Minimum length

Maximum value s Maximum length

Characters PSX abceipr s

IdNum Word number

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Appendix

IdNumLemma Lemma number

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Inf Inflectional feature infinitive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

OrthoCnt Number of spellings

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

OrthoNum Spelling number

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

OrthoStatus Status of spelling

Typ e character Null values

Minimum value A Minimum length

Maximum value B Maximum length

Characters AB

Column descriptions for English wordforms E

Part Inflectional feature participle

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Past Inflectional feature past tense

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

PhonCLX Phon wordform CELEX charset

Typ e character Null values

Minimum value NSs Minimum length

Maximum value zuz Maximum length

Characters ADEINO STUVZabd ef

ghijkl mnprstu vwxz

PhonCnt Wordform number of phonemes

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Appendix

PhonCPA Phon wordform CPA charset

Typ e character Null values

Minimum value Minimum length

Maximum value zuz Maximum length

Characters ADEIJNO STUZabd ef

ghijkl mnoprst uvwxz

PhonCV Wordform phon CV pattern

Typ e character Null values

Minimum value C Minimum length

Maximum value VVCCC Maximum length

Characters CSV

PhonCVBr Wordform phon CV pattern with brackets

Typ e character Null values

Minimum value CCCC Minimum length

Maximum value VVC Maximum length

Characters CSV

PhonDISC Phon wordform DISC charset

Typ e character Null values

Minimum value Minimum length

Maximum value tts Maximum length

Characters CDEFHIJN PQ

RSTUVZ bcdfgh ijklmnpq rs tuvwxz

Column descriptions for English wordforms E

PhonSAM Phon wordform SAMPA charset

Typ e character Null values

Minimum value Dz Minimum length

Maximum value nZeInjuz Maximum length

Characters ADEINOQ STUVZabd ef

ghijkl mnprstu vwxz

PhonStrsCLX Syll phon wordform with stress CELEX charset

Typ e character Null values

Minimum value blISnIst Minimum length

Maximum value zUOldZIsts Maximum length

Characters ADEI NOSTUVZa bd

efghij klmnprs tuvwxz

PhonStrsCPA Syll phon wordform with stress CPA charset

Typ e character Null values

Minimum value mOrl Minimum length

Maximum value zUOlJIsts Maximum length

Characters ADEIJ NOSTUZa bd

efghij klmnopr stuvwxz

PhonStrsDISC Syll phon wordform with stress DISC charset

Typ e character Null values

Minimum value mEn Minimum length

Maximum value ntts Maximum length

Characters CDEFH IJ

NPQRST UVZbcd fghijklm np qrstuv wxz

Appendix

PhonStrsSAM Syll phon wordform with stress SAMPA charset

Typ e character Null values

Minimum value TI Minimum length

Maximum value zbEstQs Maximum length

Characters ADEIN OQSTUVZa bd

efghij klmnprs tuvwxz

PhonSylCLX Syll phon wordform CELEX charset

Typ e character Null values

Minimum value SI Minimum length

Maximum value zuz Maximum length

Characters ADEINO STUVZabd ef

ghijkl mnprstu vwxz

PhonSylBCLX Syll phon wordform CELEX charset brackets

Typ e character Null values

Minimum value NSs Minimum length

Maximum value zuz Maximum length

Characters ADEINOS TUVZab de

fghijk lmnprst uvwxz

PhonSylCPA Syll phon wordform CPA charset

Typ e character Null values

Minimum value Minimum length

Maximum value zuz Maximum length

Characters ADEIJNO STUZabd ef ghijkl mnoprst uvwxz

Column descriptions for English wordforms E

PhonSylDISC Syll phon wordform DISC charset

Typ e character Null values

Minimum value Minimum length

Maximum value ntts Maximum length

Characters CDEFHIJ NP

QRSTUV Zbcdfg hijklmnp qr

stuvwx z

PhonSylSAM Syll phon wordform SAMPA charset

Typ e character Null values

Minimum value TI Minimum length

Maximum value nZeInjuz Maximum length

Characters ADEINOQ STUVZabd ef

ghijkl mnprstu vwxz

Plu Inflectional feature plural

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Pos Inflectional feature positive

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Appendix

Pres Inflectional feature present tense

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

PronCnt Number of pronunciations

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

PronNum Pronunciation number

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

PronStatus Status of pronunciation

Typ e character Null values

Minimum value P Minimum length

Maximum value S Maximum length

Characters PS

Rare Inflectional feature rare form

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Column descriptions for English wordforms E

Sin Inflectional feature st person verb

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Sin Inflectional feature nd person verb

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Sin Inflectional feature rd person verb

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

Sing Inflectional feature singular

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

StrsPat Wordform stress pattern

Typ e character Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Appendix

Sup Inflectional feature superlative

Typ e character Null values

Minimum value N Minimum length

Maximum value Y Maximum length

Characters NY

SylCnt Wordform number of phonetic syllables

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

TransIn Inflectional transformation

Typ e character Null values

Minimum value Minimum length

Maximum value yiest Maximum length

Characters bd efgiklm nprstvyz

Word Word

Typ e character Null values

Minimum value d Minimum length

Maximum value zoos Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs tuvwxy z

Column descriptions for English wordforms E

WordCnt Word number of letters

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

WordDia Word diacritics

Typ e character Null values

Minimum value d Minimum length

Maximum value epees Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs

tuvwxy zaaceeenou

WordLow Word lowercase alphabetical

Typ e character Null values

Minimum value a Minimum length

Maximum value zoos Maximum length

Characters abcdef ghijklm nopqrstu vw

xyz

WordLowSort Word lowercase alphabetical sorted

Typ e character Null values

Minimum value a Minimum length

Maximum value ttuu Maximum length

Characters abcdef ghijklm nopqrstu vw xyz

Appendix

WordRev Word reversed

Typ e character Null values

Minimum value oht Minimum length

Maximum value zzuf Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs

tuvwxy z

WordSyl Word syllabified

Typ e character Null values

Minimum value d Minimum length

Maximum value zoos Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs

tuvwxy z

WordSylCnt Word number of orthographic syllables

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

WordSylDia Word syllabified diacritics

Typ e character Null values

Minimum value d Minimum length

Maximum value epees Maximum length

Characters ABC DEFGHIJ KLMNOPQR ST

UVWYab cdefghi jklmnopq rs

tuvwxy zaaceeenou

ENGLISH COBUILD CORPUS TYPES E

Orthography Graphemic transcription Typ e

COBUILD all sources Absolute frequency Freq

Written frequency FreqW

British English written frequency FreqWB

COBUILD written sources

American English written frequency FreqWA

Undetermined origin written frequency FreqWU

Sp oken frequency FreqS

COBUILD sp oken sources British English sp oken frequency FreqSB

Undetermined origin sp oken frequency FreqSU

Appendix

Freq Absolute frequency

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

FreqS COBUILD spoken frequency m

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

FreqSB COBUILD spoken frequency British English

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

FreqSU COBUILD spoken frequency undetermined origin

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

FreqW COBUILD written frequency m

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Column descriptions for English corpus typ es E

FreqWA COBUILD written frequency American English

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

FreqWB COBUILD written frequency British English

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

FreqWU COBUILD written frequency undetermined origin

Typ e numeric Null values

Minimum value Minimum length

Maximum value Maximum length

Characters

Typ e Graphemic transcription

Typ e character Null values

Minimum value mm Minimum length

Maximum value zzzzzzrrrrr Maximum length

Characters abcdefg hi jklmno pqrstuv wxyz

COMPUTER PHONETIC

CHARACTER CODES

The tables in this app endix exemplify the disc character set in

full disc is the character set whichgives a single unique co de

to each phonetic segment in the standard sounds systems of

Dutch English and German Here segment means consonant

aricate syllabic consonant short vowel long vowel diphthong

or nasalized vowel

Each table gives the ipa characters at the far left hand side and

the corresp onding disc characters on the far right hand side In

between come examples where they o ccur of words in Dutch

English and German which exemplify the segments in question

and the co de or co des used to represent those segments in the

other character co ding sets available sampa celex and cpa

This means you can use this app endix b oth as a full overview

of disc and a checkonevery phonetic character co de used in

the celex databases If you just want to see the co des used for

one particular language then you should consult the Phonology

section of the appropriate Linguistic Guide you can also nd

general descriptions of the character sets there

Appendix

ipa Dutch English German sampa celex cpa disc

p put pat Pakt p p p p

b bad bad Bad b b b b

t tak tack Tag t t t t

d dak dad dann d d d d

k kat cad kalt k k k k

goal game Gast g g g g

lang bang Klang N N N N

m mat mad Ma m m m m

n nat nat Naht n n n n

l lat lad Last l l l l

r J rat laterrat Ratte r r r r

f fiets fat falsch f f f f

v vat vat Welt v v v v

S thin T T T T

then D D D D

s sap sap Gas s s s s

z zat zap Supp e z z z z

M sjaal sheep Schi S S S S

ravage measure Genie Z Z Z Z

j jas yank Jacke j j j j

xc licht gaat loch Bachich x x x x

regen G G G G

h had had Hand h h h h

w why waterpro of w w w w

Y wat w w w w

pf Pferd pf pf pf

ts Zahl ts ts C

Q cheap Matsch tS tS T J

jazz jeep Gin dZ dZ J

bacon N N N C

j

m idealism m m m F

j

n burden n n n H

j

l dangle l l l P

j

father linking r r r r R

Disc Computer Phonetic Codes

consonants affricates and syllabic consonants

Computer phonetic character co des

ipa Dutch English German sampa celex cpa disc

iq liep bean Lied i i i i

iqq analyse i i i

q barn Advantage A A A

aq laat klar a a a a

q born Allroundman O O O

uq boek boon Hut u u u u

q burn Teamwork

yq buut fu r y y y y

yqq centrifuge y y y

q scene Kase E E E

q freule U Q

q zone Q O o

eq leeg Mehl e e e e

q deuk Mobel q

oq boom Boot o o o o

e bay Native eI eI e

a buy Shylo ck aI aI a

boy Playboy OI OI o

V no U U O

aV brow Allroundsp ortler aU aU A

peer I I I

pair E E E

V poor U U U

i wijs EI EI y K

y huis I UI q L

u koud Au AU A M

ai weit ai ai a W

au Haut au au A B

y freut Oy Oy o X

Disc Computer Phonetic Codes

long vowels and diphthongs

Appendix

ipa Dutch English German sampa celex cpa disc

lip pit Mitte I I I I

Pfu tze Y Y Y Y

leg pet Bett E E E E

Gotter Q Q

pat Ragtime

a hat a a a

lat Kalevala A A A A

pot Q O O Q

putt Plumpudding V V V

bom Glocke O O O O

V put Pult U U U U

T put U Y

gelijk another Beginn

q Parfum Q Q

timbre impromptu c

q detente Detente A A A q

q lingerie Bassin

q b ouillon Aront O O O

Disc Computer Phonetic Codes

shortvowels and nasalized vowels

ipa Description sampa celex cpa disc

q length marker

syllable marker

h primary stress

i secondary stress

nasalization

examples A A A

Disc Computer Phonetic Codes

length stress syllable and nasalization markers

ASCI I AND EIGHTBIT

CHARACTER CODES

The two tables which followshow full details of the seven

and eightbitcharacter co des used by celex on its digital

vaxvms computer systems They are particularly useful when

you need to transfer data to or from the celex machine you

can nd out which co des must b e converted The rst table

shows the basic characters in use they are the standard seven

bit asci i co des and most asci i terminals and printers should

repro duce these characters as shown The second table shows

the eight bit co des which digital vt and vtseries

terminals can repro duce these are the co des which provide the

diacritic characters available in some columns in the celex

databases

Most of the printable seven and eight bit co des conform to

the standard character set known as iso Latin Alpha

b et No or ecma There are some exceptions however

The iso decimal characters

and are not imple

mented in the digital set and and each pro duce

acharacter other than the iso recommended one

For details ab out eachcharacter consult the digital vms Gen

eral User Guide Volume A Guide to using VMS VMS version April pages AA

Appendix

ENQ

NUL SOH STX ETX EOT ACK BEL

BS HT LF VT FF CR SO SI

A B C D E F

DLE DC DC DC DC NAK SYN ETB

CAN EM SUB ESC FS GS RS US

A B C D E F

SP

B F A C D E

C D E F A B

A B C D E F G

H I J K L M N O

A B C D E F

Q

P R S T U V W

X Y Z

B D A C E F

g

a c e

b d f

m n o

j

h i k l

B C D E F A

p q

r s u v w

t

y

x z

DEL

B C D E A F

A B C D E F

Octal

Character Decimal

O

F

Hexadecimal

DIGITALCELEX SEVENBIT ASCI I CODES

Ascii and eightbit character co des

IND NEL SSA ESA

HTS HTJ VTS PLD PLU RI SS SS

A B C D E F

DCS PU PU STS CCH MW SPA EPA

CSI ST OSC PM APC

A B C D E F

x

Y

C

A A A A A A A A

A A

bacd

c



a

A AB A AA AC AD AE AF

B B B B B B B B

B B

o



B B BA BC BD BE BB BF

C

A

A A A A A

C C C C C C C C

C C

E E E E I I I I

C C CA CB CC CD CE CF

N O O O O O

D D D D D D D D

D D

U U U U Y

DF D DE D DA DB DC DD

c

a a a a a a

E E E E E E E E

E E

e e e e

E E EA EB EC ED EE EF

n o o o o o

F F F F F F F F

F F

y

u u u u

FD FE FF F F FA FB FC

A B C D E F

Octal

Character Decimal

O

D

Hexadecimal

DIGITALCELEX EIGHTBIT CODES

GLOSSARY

Glossary ABBREVIATION BACKTRACK KEY

ABBREVIATION A term which refers to the shortened form of a normal word

or phrase which can b e used when the word or phrase itself is thoughttobe

to o long or unwieldy A sp ecial lexicon type whichcontains nothing but

abbreviations is available An abbreviation can take one of three general

forms of which the most common is the contraction where particular letters

often vowels are removed from the word thus Gld is an abbreviation of

Gelderland Another form is the acronym where the initial letters of each

constituentword in a phrase are joined to makeanewword thus FIFA is

an acronym of Federation Internationale de Fo otball Asso ciation Finally

there is also truncation where a numb er of letters is removed from the end

of a word thus the chemical symbol for Argon is Ar

ABSTRACT STEM This term refers to an alternative orthographic form of

the stem When a stem ends in s or f and when in any of its related

wordforms that f b ecomes a v the verb leven ik leef wij leven or

that s b ecomes a z the noun kaas singular kaas plural kazen then the

abstract stem is given with the endings v and z resp ectively thus leev

and kaaz instead of the normal stems leef and kaas All other abstract

stems have the same form as the normal stem

AFFIX SUBSTITUTION This refers to the pro cess by which an ax replaces

part of a lemma when the ax and the lemma combine to make a new

lemma An example is the English lemma fatuity where the headword

fatuous can b e said to lose the ax ous and gain the ax ity

ALPHABETIC KEYS This refers to the lettersas opp osed to the numbers

and various other symb olsthat are on your keyb oard twentysix upp er

case and twentysix lower case characters When you are working in flex

you can use them to move to the nearest menu option whichbeginswith

the letter you press

AND OPERATOR The logical connectivecombining two restrictions or

groups of restrictions x and y in sucha waythatarow is included

in the lexicon only if b oth x and y are true for that row otherwise the

row is not included in the lexicon

ASCI I The sevenbit binary numb er co ding system used to represent alphab etic

numeric punctuation and other characters in some typ es of computers The

letters stand for American Standard Co de for Information Interchange

BACKTRACK KEY This is a flex term which refers to the key you press to

reverse backdown the menu path you have just come along You generally

use the backtrackkey to leave a menu windowyou do not wish to use and

to return to the previous window

BAR COPY Glossary

BAR This refers to the way flex indicates which option you are cho osing in a

particular window Usually b old text underlined text or reverse video text

is used to dierentiate the current option the one you will get if you press

return from the others

BATCH MODE This allows you or flex working for you to submit cer

tain commands to the computer which it carries out as a separate non

interactive job flex uses batch mo de for certain typ es of job b ecause in

this way they are executed while the computer is not b eing used for smaller

jobs or more imp ortant jobs

BELL This refers to the noise your terminal can make to attract your

attention In flex the b ell normally sounds when a new window app ears

or when a message is displayed

CANCELLED This refers to the status of a flex job which either you or

flex has asked to b e stopp ed When you cancel a job the computer stops

working on it and ignores any results already achieved by that job

CLASS LABELS A simple co ding system used to indicate the syntactic class

ofaword n means a noun a means an adjective and so on They can b e

used instead of a numeric co ding system or typing the syntactic class in

full

COBUILD This is an acronym for Collins Birmingham University Internation

al Language Database In cobuild published the Collins Cobuild

English Language Dictionary which is based on analysis of their large

corpus of mo dern English The frequency information in the celex

English database was taken from this corpus which at the time contained

words

COLUMN A database term which refers to the storage of one particular typ e

of information a column can contain a sp ecic sort of words or co des or

analyses

COMPLETE SEGMENTATION This means the full derivational morphologi

cal analysis of a lemma into all its constituent morphemes

COMPLETED This is a flex term which indicates that a job working in

batch mode has now nished successfully

COPY This is a flex term which refers to the creation of a new lexicon

by using the denitions ie the columns and restrictions already

sp ecied for a dierent lexiconThelexicon you copy can b e your own

or if you havea grant someone elses

Glossary CORPUS DBMS

CORPUS A sizeable collection of words usually written texts which can

b e used and pro cessed by computers Three text corp ora were used to

provide celexs frequency information the inl cobuild and eind

hoven corp ora They all contain mo dernday texts drawn from diverse

printed sources such as recentlypublished b o oks newspap ers and maga

zines and sometimes though to a much lesser extent transcriptions made

from recordings of sp eech

CORPUS TOKEN A term which refers to the units distinguished during the

disambiguation by computer of a text corpus used to provide fre

quency information A corpus token is any string containing at least one

alphabetic character along with zero or more alphanumeric characters

The inl corpus contained tokens and the cobuild corpus

contained tokens

CORPUS TYPE A term which refers to a corpus token that o ccurs one

or more times in a corpus During the pro cess of disambiguation the

o ccurrence corpus tokens can b e quantied Whenever a new corpus

token is discovered it is also noted as a corpus typ e and thereafter any

reo ccurrences counted to givethe frequency count of the typ e The typ e

which accounts for the greatest number of tokens in the inl corpus is de

it o ccurs times

CPA A computer phonetic alphab et develop ed the Ruhr Universitat Bo chum

The letters stand for Computer Phonetic Alphab et

CURSOR An indicator usually a small ashing b ox or a line used to indicate

where the next character will app ear or in flex to mark the currentmenu

option usually in conjunction with a bar

CV PATTERNS A cv pattern is a rewritten orthographic phonetic or phono

logical transcription in which generally sp eaking anyvowels or diphthongs

are replaced by the letter v and consonants by the letter c

DATABASE A database is a collection of information stored in computer les

in sucha wayastomake the retrieval of that information quicker and more

exible

DATANET This is the name of the main public psdn in the Netherlands

At present all surfnet no des are datanet no des since surfnet uses

datanet

DBMS These letters stand for database management system which is computer

software designed to facilitate the use and developmentofa database

celex and flex use the relational dbms marketed bytheoracle company

DELIMITER EXPRESSION Glossary

DELIMITER This refers to a character or group of characters used in a file

to indicate the b eginning or end of every field

DERIVATIONALCOMPOSITIONAL SEGMENTATION This is the the typ e

of morphological analysis which identies the constituent lemmasaxes

and morphemes in a lemma as opp osed to inectional analysis whichdeals

with the wordforms each lemma takes

DIACRITICS The markers used in conjunction with regular orthographic char

acters to indicate some dierence in pronunciation or stress as with the

Germanumlaut the Frenchacute and the Czechhacek

DISAMBIGUATION This term refers to the pro cess by whichthe frequency

of words in a large text corpus can b e established either by computer

or p eople or b oth The pro cess tries to link eachword in the corpus that

is each string consisting of one alphanumeric charcter plus at least one

alphab etic character with a space on either side with a lemma If a string

o ccurs more than once and if such a link can b e made then the word is

considered to b e a wordform and the numb er of times the link was made

is the frequency of that wordform

DISC This is the name of the celex computer phonetic alphab et whichusesone

unique distinct character for eachvowel long vowel diphthong consonant

and aricate Although not elegant in app earance it is useful for computer

pro cessing

DRAFT A flex term which refers to the version of a lexicon It indicates

that its denition is stored by flex and that when you use the lexicon

the information is extracted from the main celex database using that

denition Contrast with fixed

DTE These letters stand for data terminal equipment which for most celex

users normally just means computer

ETHERNET A sp ecial communications setup for a lan which allows dierent

sorts of computers and other devices to b e linked without central control

from any one computer

EXECUTING This is a flex term which indicates that a job is currently b eing

carried out in batch mode

EXPORT This is a flex term which refers to the pro cess of making a normal

vaxvms le from the contents of a lexicon

EXPRESSION This is a flex term which refers to the righthand part of a

restriction that is the part whichcontains some number word or wild

card A column name is linked to an expression by means of an operator

Glossary FIELD GATEWAY

FIELD In flex this refers to that part of a window where information from

the database app ears In a vaxvms file it refers to a sp ecic part of a

line which is used for a particular sort of information

FILE A collection of data stored for computer use and arranged in a way which

is signicant to the user

FINITE FORMS This refers to those ections which can o ccur in their own

right in a main clause or sentence and which indicate dierences in tense

and p erson for example ik b eweg ik b ewog wij b ewegen wij b ewogen

FIXED A flex term which refers to the version of a lexicon A fixed

lexicon is a separate indep endent database whichcontains information

originally taken from the central celex databases and when you use it the

information is extracted from this database rather than the central celex

database Contrast with draft

FIXED FORMATFILE This is a file whose fields are always a xed number

of characters wide regardless of the width of the data each eld contains

FLAT SEGMENTATION This is one typ e of derivationalcomp ositional mor

phological analysis It reduces a lemma directly to its constituentmor

phemes without showing any of the intermediate levels of analysis you get

Contrast with hierarchical segmentation

FLEXEXP This is a logical name which refers to the directory of your

celex account which is set aside sp ecically for files which are extracted

from flex using the export facility

FREQUENCY The numb er of times a corpus type o ccurs in a particular

corpusFor example the wordform radio has an inl frequency of

as counted in the word inl corpus This gure can also b e

expressed prop ortionally ie the frequency exp ected p er million words

or logarithmicallyTo arrive at a gure for the frequency of lemmas the

frequencies of its inectional forms that is its wordforms are added

together

FULLY SYLLABIFIED This refers to orthographic transcriptions whichhave

a syllable marker whenever a syllable b oundary o ccurs within a word

including singleletter syllables which o ccur at the b eginning or end of a

word Contrast with partially syllabified

GATEWAY This refers to the p ointofinterconnection b etween two dierent

communications networks Often users are not aware they are using gate

ways o ccasionally though you may rst have to connect to a gateway

b efore b eing able to use the other network

GRANT INL Glossary

GRANT This is a flex term that indicates whether one or more particular

flex users or every flex user can copy a lexicon created byyou

GRAPHEMIC This is the adjectiveusedtodenotecharacters which o ccur in

normal Dutch English or German orthography It is used to distinguish

phonetic or phonological transcriptions which use sp ecically phonetic

character alphab ets from transcriptions which are written or typ ed using

the roman alphab et

HEADWORD Atermwhich refers to one of the two forms a lemma is given

in the celex databases It corresp onds to the traditional lexicographic

headword to b e found in dictionaries In Dutch German and English the

forms used always resemble words that o ccur naturally in the language

rather than abstract forms Thus in Dutch the headword of a noun is its

singular form For a denitive list of the forms used consult App endix iv

Contrast with stem

HELP KEY This is a flex term that refers to the key you press to receive

online advice on howtouseflex as you are working with it

HIDDEN This is a flex term which refers to the columns displayed using

the show option If your lexicon contains so many columns that not

all of them can b e displayed at once on screen then you can indicate that

certain columns should temp orarily b e missed out of the display so that

you can see other columns of more interest The missed out columns are

called hidden columns

HIERARCHICAL SEGMENTATION This is one typ e of derivational or com

p ositional morphological analysis It reduces a lemma directly to its con

stituent morphemes showing all the intermediate levels of analysis involved

in arriving at all the morphemes Contrast with flat segmentation

IMMEDIATE SEGMENTATION This is one typ e of derivational or com

p ositional morphological analysis It reduces a lemma to its next biggest

comp onents other lemmas axes or morphemes To arriveat complete

segmentation immediate segmentation mayhave to b e carried out

several times

INDEX This is a database term which refers to columns whose contents are

indexed in a way conceptually identical to the indexing of b o ok Information

from columns with an index can b e lo oked up more quickly bythedbms

INL This is the normal abbreviation for Instituut voor Nederlandse Lexicologie

the Dutch Lexicography Institute in Leiden They are developing a large

text corpus of mo dern written Dutch and the frequency information

contained in the celex Dutch database was extracted from this corpus

when it contained over million words It is still b eing extended and now

contains more than million words

Glossary INTEGRITY LEVEL

INTEGRITY A term which refers to the protection of information stored in

a database when it can b e altered bytwo or more sources A database

maintains its integrity so long as only one source can alter the data at

any one time If two p eople try to alter the same data at the same time

the resulting information is no longer consistent and the integrity of the

database is lost

INTERVAL This is a flex option which allows you to sp ecify a particular set

of consecutive rows in your lexicon for export

IPA This letters stand for International Phonetic Alphab et the set of written

characters approved for phonetic transcription by the International Pho

netic Asso ciation

ISO These letters stand for the International Standards Organization the

Swissbased organization whichisinvolved in developing and co ordinating

worldwide standards

LAN These letters stand for lo cal area network and refer to a communications

network whichlinksanumb er of computers over a relatively small area

such as a factory plant or university

LAT These letters stand for lo cal area transp ort and refer to the proto cols a

dec terminal server uses to communicate with computers using vaxvms

over an ethernet

LANGUAGE CODES These are co des used in the English database to provide

background information ab out some lemmas such as the national origin

words loaned from other languages and whether certain lemmas are more

likely to b e British or American English

LEMMA A term intended to signify the abstract notion which underlies a family

of inected forms so that for example walk could b e the lemma underlying

the verbal forms walk walks walked and walking In the celex databases

lemmas are distinguished on the basis of the pronunciation the

syntactic class the morphological structure the orthographic form

of their various wordformsaswell as the full inectional paradigm of

the lemma No explicit consideration of meaning is involvedsoin

the celex databases the lemmas of anytwo or more words whichdier

in meaning but which otherwise are identical in each of these veways are

reduced to one lemma In principle anyconvenient form could b e used to

represent a lemma an abstract form or even a numb er In practice celex

uses two forms the headword and the stem

LEVEL This refers to any one analytical step in morphological analysis com

plete segmentation is nished when every p ossible level of analysis has b een carried out

LEXICON MENU Glossary

LEXICON This term refers to a subset of one of the celex databases whichyou

can dene for yourself using flex Rather than using the entire database

at all times you sp ecify certain columns and delimit their contents using

restrictions to form a coherent subset of information drawn from the

central database

LEXICON TYPE This is a flex term that indicates which of the central celex

databases the information in a lexicon is drawn from Each of the central

databases has as its main sub ject one typ e of canonical form such as Dutch

lemmas or English wordforms The typ e of canonical form is then used to

indicate the typ e of lexicon

LISP This is a high level programming language often used in articial intel

ligence work In particular it uses a sp ecial brackets notation for its input

and output data

LOCKED This means that flex is currently working on your lexicon and

that in order to protect its integrityyou cannot do anymorework with

it until the job flex is doing has nished

LOGICAL COMBINATION Thisisa flex term which refers to the way

restrictions or groups of restrictions linked by brackets work together

to delimit the contents of a lexicon by means of the AND operator

the OR operator and the NOT operator

LOGICAL NAME A vaxvms term which refers to a sp ecic directory in

your account It is used as part of a file name to help you to remember

where it is and the computer to knowhow to nd it or store it

LOGIN This refers to the wayyou identify yourself to the computer b efore

b eginning anywork You normally havetogive the name of your account

and a password

LOGOUT This refers to the wayyou indicate to the computer that you want

to stop working On the celex machine you simply typ e logout

MAIL This a flex term and a vms term In flex it refers to the main menu

option mail whichallows you to communicate with other flex users purely

within flex it do es not link in with the other national or international

networks In vms there is a more comprehensive mail facility which allows

you to send messages to other celex computer users as well as users on

other computers via decnet or datanet

MENU This is a flex term whichreferstotheboxes displayed on your screen

from whichyou can cho ose an option that allows you to continue with your

work or a particular typ e of information Compare window

Glossary MESSAGE LINE OR OPERATOR

MESSAGE LINE This is a flex term which refers to the line immediately

ab ove the b ottom line of the screen It displays instructions error messages

and other information to help you as you use flex and whenever celex

computer system messages are senttoyour terminal they are also displayed

here

MODEM This is an an acronym for the words mo dulator and demo dulatorIt

is a machine whichconverts the characters from your computer a digital

bit stream into a form an analog signal that can b e transmitted along

a telephone line this is mo dulation It can also convert the analog signal

received down a telephone line backinto the digital bit stream used in your

computer this is demo dulationThus you can use telephone lines to work

interactively with a computer that mightbelocatedhundreds of miles away

provided that you have a terminal and a mo dem and the remote computer

is also linked to a mo dem

NEXT KEY This is a flex term which refers to the key you press to display

more information in a window or menu

NOT OPERATOR The logical connective applied to one restriction or group

of restrictions z in sucha way that a row is included in the lexicon if

z is untrue If z is true the row is not included in the lexicon

ON VIEW This is a flex term which is imp ortant for columns that are used

in the construction of restrictions If a column is on viewyou can see

it when you displayyour lexicon using the show or export options If it

is not on viewyou never see it but it still works in any restriction you

have made with it All columns are on view by default you can change

this in the edit restrictions menu

OPERATING SYSTEM This refers to the software whichyou use sp ecically

to control a computer or a computer system The commands you typ e to

start a program running or to givea directory listing are operating

system commands The celex computers use vaxvms

OPERATOR This is a flex term which refers to the simple mathematical

relation symb ols that you can use in restrictions

OR OPERATOR The logical connectivecombining two restrictions or

groups of restrictions x and y in sucha way that a row is included in

the lexicon if i either x or y is true for that row or ii b oth x and y

are true for that row otherwise the row is not included in the lexicon

PAD QUERY Glossary

PAD These letters stand for packet assemblerdisassembler a device or pro

gram which gathers individual characters that you send from your ter

minal or computer and puts them into groups that is packets which can

then b e sent across a psdn to some other computer Likewise when packets

come backtoyour computer across the psdnthepad splits them up into

individual characters again ready for displayonyour terminal

PAGE This is a flex term that refers to data displayed in the show window

there is ro om for ten lines of information on screen and one page is equal

to these ten lines

PARTIALLY SYLLABIFIED This refers to orthographic transcriptions which

indicate each syllable b oundary within a word by means of a hyphen with

the exception of syllables at the b eginning or end of the word which consist

of only one letter such syllables are not marked Compare with fully

syllabified

PENDING This is a flex term which means that a batch job cannot b e

executed by the computer at the moment usually b ecause other batch

jobs are b eing carried out A job whichis pending will eventually b e

carried out however unless you cancel it

PREV KEY This is a flex term which refers to the key you press to redisplay

old information that you have already seen in the window or menu you

are currently working in

PSDN These letters stand for packet switching data network which is a wide

area network that can control the rapid transmission of packets of data

p ossibly prepared byapad for example b etween dierent p oints in the

network psdns enable you to work interactively on a computer whichis

lo cated hundreds of miles away In the Netherlands the public x psdn

is called datanet and it is currently used in the implementation of

surfnet

PSI These letters stand for packetnet system interfacethevaxvms software

pro duct that enables vax computers to link up with psdns It p erforms the

function of a pad

PSS These letters stand for packet switchstream the name of the British x

psdn

QUERY This is a flex term that refers to the show menu option that allows

you to lo ok at a particular part of your lexicon It do es not p ermanently

alter your lexicon

Glossary REDRAW STRESS PATTERN

REDRAW This is a flex term that refers to the key whichyou press to re

display all the flex information currently displayed on screen It allows

you to correct any badlydrawn lines or get rid of unwanted messages or

straycharacters

RESTRICTION This is a flex term which refers to a simple logical statement

you formulate to sp ecify in detail the information to b e included in your

lexicon with reference to the contents of the columns already in your

lexicon

ROW A database term which refers to the storage of dierenttyp es of infor

mation which refer to one word eachrowcontains an orthographic tran

scription a phonetic transcription a morphological analysis a syntactic

co de and a frequency count and more b esides for eachword

SAMPA These letters stand for Sp eech Assessment Metho ds Phonetic Al

phab et sam is an Esprit Europ ean Community funded pro ject The

development of the phonetic alphab et was coordinated byJohnWells with

the intention of it b ecoming the standard Europ ean computer phonetic

alphab et

SEGMENTATION This is a term which refers to the pro cess of morphological

analysis of words into their constituent lemmas axes and morphemes

SQLPLUS This is the name of the standard dbms pro duced by the ora

cle companyItisdbms used by celex when you work with flexyou

are using a system which generates sqlplus co de to access the celex

databases

STATUS LINE This is a flex term which refers to the very b ottom line of the

screen It displays your flex username the name of the lexicon you have

selected and version number of the flex program you are using

STEM A term which refers to one of the twoformsalemma is given in

the celex Dutch database and the term used in place of headword in

English morphology It is that part of a lemmas inectional paradigm

which is common to all the inected forms separate from the inectional

axes themselves Usually it is identical to the headword except for

Dutchverbs where it takes the form of the rst p erson singular present

tense but see also abstract stem In English morphologyastem is a

headword or sometimes a ectional form of a headword

STRESS PATTERN This refers to sp ecial strings of numb ers each of which

represents one phonetic syllable and indicates how that syllable is stressed

A zero always means unstressed a indicates stressed in stress pat

terns for Dutchwords and primary stress for English words indicates

secondary stress for English words

SURFNET WAN Glossary

SURFNET This is the Dutch national academic computer network which pro

vides electronic mail facilities and logins to computers all over the Nether

lands Atpresent it uses datanet to carry out its work

SYLLABIC CONSONANT This term refers to a consonant whichby itself or

with other consonants forms a distinct syllable in the pronunciation of a

word without the presence of a vowel The nal l in the word b ottle can

b e realised as a syllabic consonant

TERMINAL This is a device which can accept data from and transmit data to

a computer For most p eople a terminal is a visual display unit vdu for

short which consists of a televisionlike screen to display data received

andakeyb oard to transmit data including operating system commands

There are manytyp es of terminal all with their own sp ecic control co des

and capabilities

TERMINAL EMULATOR Thisisatyp e of software which allows your p ersonal

computer or terminal to b ehave and resp ond like another sort of terminal

TERMINAL SERVER A device that connects terminals and mo dems and

printers to an ethernet

VAXVMS The trademark used by the Digital Electronic Corp oration dec

to identify the operating system used on their vax series computers

vax stands for Virtual Addressing eXtension and vms stands for Virtual

Memory System

VERSION This is a flex term that refers to the wayyour lexicon is stored If

it is a draft lexicon only the denition is stored and when you use it the

data it requires is lo oked up in the main celex database If it is a xed

lexicon it is a separate probably much smaller database which is quicker

and easier used

VT This refers to a standard dec typ e of terminalUserswhohave sucha

terminal or who havea terminal emulator which can imitate sucha

terminal should b e able to log into celex and use flex with no problems

VT This refers to a standard dec typ e of terminal which is newer than

the vt It is the default terminal typ e for celex and flex

WAN These letters stand for wide area network and refer to a communications

network which links a numb er of computers over a relatively large area

Sometimes these networks cover entire nations suchas surfnet in the

Netherlands or even larger areas suchas earn the Europ ean academic

network

Glossary WILDCARD X

WILDCARD This refers to the and characters which can b e used in a

restriction or query to indicate resp ectively anycharacter or group of

characters and any single character

WINDOW This is a flex term which refers to the b oxes shown on your

screen whichcontain either menu options data drawn from the database

or other relevant flex information A window whichcontains options is

almost always called simply a menu

WORDFORM A term which is synonymous with word in the general sense

Wordforms are the units o ccurring in natural language which when writ

ten are b ounded on either side by a space and which can b e asso ciated

with a lemmaHowever some English and Dutchwordforms include spaces

swimming p o ol for example or Nederlandse Sp o orwegenTheyare

the inflected forms in regular use as opp osed to lemmas stems

and headwords which are convenient but abstract representations of

complete families of wordforms

X This refers to the standard proto cols recommended by the Comite Consul

tatif International Telegraphique et Telephonique for equipment op erating

within a psdn

X This refers to the standard pro cedures recommended by the Comite

Consultatif International Telegraphique et Telephonique for the exchange

of user data and the required control information b etween your terminal

and a remote pad over a psdn