?tekauer Pavol, Lieber Rochelle. Handbook of Word-Formation

Подождите немного. Документ загружается.

LEXICALIZATION AND INSTIT

TIONALIZATION

371

(

e.g. Lieber 1988, 1992, Bauer 1983) or by is

lating the scope of this apparently

yntactic element within word-formation proper (cf. Hohenhaus 1996: 218-228)

particularly interesting proposal is the integrational approach of Jackendoff

(

1997

)

. The em

irical basis for his endeavour comes from his ‘Wheel o

Fortune

us’ – a collection

(

com

iled over a f

w months

)

of ca. 600 fixed ex

ressions

used in the

ame show of the same name, where contestants have to solve

uzzles as

uickly as possible from the step-by-step disclosure of letters. What is striking is

hat there seems to be little strain (on the part of the programme makers) in coming

up with and (on the part of the contestants) recognizing vast amounts of such strings

(

and the show has been on air for many years six days a week). Strings in question

nclude not only idioms proper such as

to throw in the towel

to eat humble pie

t al

clichés

suc

h a

we’re doing everything humanely possible

imme a

eak

and well-known

uotations such as ma

the Force be with

beam me

up, Scott

. Quite obviousl

, then, “there are too man

idioms and other fixed

expressions for us to simpl

disre

ard them as phenomena on the mar

in of

grammar” (Jackendoff 1997: 177), and it is even possible that “their number is of the

ame order of magnitude as the single words of the vocabulary” (Jackendoff 1997:

156). The reason that such fixed expressions have typically been neglected esp. in

generative theories (as ‘not core g

ammar’) probably does stem, as Jackendoff

(

1997: 153, 157) surmises, from the simple conception of the lexicon as the source

s at the point of lexical insertion onl

.Havin

replaced that notion (see section

2.3

)

with a model of tri

artite

arallel

exical licensin

, Jac

endoff (1997) mana

o inte

rate listemes of all formats neatl

Suffice it to say here, by way of a conclusion: it is clear that the lexicon has to be

ore than a simple list of ‘words’. Rote-learning, memorized building blocks of

arious sizes and associations between them, form a large and integral part of

‘lexical knowledge’ alongside (competence-)knowledge of morphemes plus the

productive morphological rule-system fo

on-line (de-)composition of complex

words. In short: lexicalization/listing is of great relevance even beyond word-

rmati

ete

r H

ohenhaus

University of Nottingham

epartment of German

rent Building

University Par

University of Nottingham

Nottingham NG7 2RD, UK

-mail: Peter.Hohenhaus

nottin

ham.ac.u

More recent evidence of the im

ortance of larger-than-word memorized building-blocks of language

also comes from Wra

(2002).

2 PETER H

HENHA

EFEREN

itchison

Jean. 1994.

ords in the Mind – an introduction to the mental lexicon.

edition

Oxford:

ronoff

Mark. 1976.

ord Formation in

enerative

rammar

Cambrid

e, Mass.: MIT Press.

ronoff

Mark. 1988. Review o

Di Sciullo/Williams

(

1987

)

. Language 64

766-770.

ronoff, Mark and Anshen, Frank. 1998. “Morpholo

and the Lexicon.” In: A. Spencer and A. Zwick

(

eds.

)

, 237-247.

auer

Laurie. 1983.

nglish Word-formation. Cambridge: Cambridge University Press.

auer

Laurie. 2001.

orphological Productivit

Cambridge: Cambridge University Press.

rekle, Herbert and Kastovsky, Dieter (eds.). 1977.

erspektiven der Wortbildungsforschung. Beiträge

um Wuppertaler Wortbildungskolloquium vom 9. – 10 Juli 1976. Anläßlich des 70. Geburtstages

von

ans

chand am

ktober 1977.

ouv

oseriu, Eugenio. 1967. “Lexikalische Solidaritäten.”

Poetica

1, 293

203

stal, David. 2001. Lan

e and the Interne

. Cambrid

e: Cambrid

e Universit

Press.

e Jon

, Niv

a H.; Schreuder, Robert; and Baa

Harald, R. 2000. “The morpholo

ical famil

size

ffect and morpholo

.” Lan

e and Co

nitive Processes

, 329

Di Sciullo

Anna-Maria and Williams

Edwin. 1987.

On the Definition of Word

Cambridge, Mass.: MIT

ess.

Downin

, Pamela. 1977. “On the creation and use of En

lish compound nouns.” Lan

810-842.

Halle, Morris. 1973. “Prolegomena to a theory of word-formation.” Linguistic Inquir

3-16.

Herbermann, Clemens-Peter. 1981. Wort

Basis

Lexem und die Grenze zwischen Lexikon und

rammatik. M

Fink

Heringer, Hans Jürgen. 1984. “Wortbildung: Sinn aus dem Chaos.” Deutsche S

rach

1-13.

Hohenhaus

Peter. 1996. Ad-hoc Wortbildung – Terminologie

Typologie und Theorie kreativer

Wortbildung im Englischen

Frankfurt/M., Berlin, Bern, New York, Paris, Wien: Peter Lang.

Hohenhaus, Peter. 1998. “Non-Lexicalizability as a Characteristic Feature of Nonce

ord-Formation in

lish and German.” Lexicolo

4/2

237-280.

Hohenhaus

Peter. 2000. “An Ove

ooked T

pe of Word-Formation: Dumm

-compounds in German and

lish.” In: Ch. Hall and D. Rock (eds.),

erman

tudies towards the Millennium. Oxford

Bern

erlin

Bruxelles

Frankfurt

New York

Wien

241-260.

Hohenhaus

Peter. 2004. “Identical Constitu

nt Compounds – a Corpus-based Study.” Folia Linguistica

38/3-4

297-331.

Hohenhaus

Peter. 2005a. “Elements o

traditional and ‘reverse’ purism in

relation to computer-mediated

ommunication.” In: N. Langer and W. Davies (eds.), Linguistic Purism in the Germanic Languages

erlin, New York: De Gruyter, 186-202.

ohenhaus

Peter. 2005b. “Where Word-Formati

n cannot Extend the Vocabular

- Creativit

Productivit

and the Lexicon in S

nchronic and Diachronic Morpholo

.” In: D. Kastovsk

and A.

Mettinger (eds.), Lexical Change and the Genesis of the English Vocabular

Berlin

New York:

Mouton de Gru

ter.

hes, Geoffre

. 2000.

Histor

lish Words

xford: Blackwell.

Jackendoff, Ray. 1975. “Mor

phological and Semantic Regul

ariti

in th

” Lan

639-

671.

ackendoff, Ray. 1997.

he Architecture of the Language Facult

Cambridge, Mass.: MIT Press.

Kastovsky, Dieter. 1982a.

Semantik und Wortbildun

. Düsseldorf

Bern

München: Ba

el/Francke.

Kastovsky, Dieter. 1982b. “Word-Fo

mation: A Functional View.” Folia Linguistica 16, 181-198.

Kastovsky, Dieter. 1993.

he Place of WF in Grammar: A Historical Survey

ICOLS.

Kjellmer, Göran. 2000. “Potential Words.”

Word

51/2, 205-228.

Leech, Geoffrey. 1974.

emantics. Harmondsworth: Penguin.

Lehmann

Christian. 2002. New reflections on grammaticalization and lexicalization

In: I.

ischer and

. Diewald

(

eds.

)

, New Reflections on Grammaticalization

Amsterdam, Philadelphia: John

enjamins, 1-18.

Lieber

Rochelle. 1981.

n the Or

anization o

the Lexicon. Bloomin

ton, Indiana: Indiana Universit

Lin

uistics Club.

Lieber, Rochelle. 1988. “Phrasal Compounds in En

lish and the Morpholo

-S

ntax Interface.” In:

Papers from the 2

Annual Regional Meeting of the Ch

cago Linguistics Society. Part Two:

Parasession on Agreement in Grammatical Theory

)

, 175-184.

EXICALIZATION AND INSTIT

TIONALIZATION

373

Lieber

Rochelle. 1992. Deconstructing Morphology.

hicago, London: University of Chicago Press.

Lipka, Leonhard. 1977. “Lexikalisierun

, Idioma

isierun

und H

postatisierun

als Probleme eine

nchronen Wortbildun

slehre.” In: H. Brekle and D. Kastovsk

(eds.),1977, 155-163.

Lipka, Leonhard. 1981. “Zur Lexikalisi

im Deutschen und En

lischen.” In: L. Lipka and H. Günthe

(

eds.

)

, Wortbildun

Darmstadt:

issenschaftlich

Buch

esellschaft, 119-132.

ka, Leonhard. 2002.

lish Lexicolo

: Lexical structure, word semantics & word-

ormation

Tübin

en: Narr.

Lyons, John. 1977. Semantics

Cambridge: Cambridge University Press.

McQueen, James M. and Cutler, Anne. 1998. “Morphology

n Word Recognition.” In: A. Spencer and A.

wicky (eds.),1998, 406-427.

McWhorter

John. 2001. The Power of Babel – a Natural History of Languag

New York: Henry

Tim

es.

Meys, Willem, 1975. Compound Adjectives in English and the Ideal Speaker-Listener. Am

rth

llan

Meys, Willem. 1985. “Morphological meaning and the struct

ure of the mental lexicon.” In: A. J. Geer, P.

oppenbrouwers, A. M. Seure

and T. Wei

ters (eds.),

eanin

and the Lexicon

Dordrecht: Foris

76–85.

Motsch, Wolf

. 1977. “Ein Plädo

er für die

eschreibun

von Wortbildun

en auf der Grundla

e des

Lexikons.” In: H. Brekle and D. Kastovsk

(eds.), 180-202.

Pinker, Steven. 1994.

he Language Instinct. London: Penguin.

Selkirk

Elizabeth. 1982. The Syntax of Words. Cambridge, Mass.: MIT Press.

Spencer, Andrew and Zwicky, Arnold. 1998. The Handbook of Morpholog

Oxford: Blackwell.

Štekauer

Pavol. 2000.

nglish Word-Formation – A History of Research (1960-1995

)

. Tübingen: Gunte

arr

Štekauer, Pavol. 2002. “On the Theory of

Neologisms and Nonce-formations.”

ustralian Journal o

Linguistics 22/1

97-112.

Toman

JindĜ

1983. Worts

nta

. Tübin

en: Nieme

er.

einreich, Uriel. 1969. “Problems in the Anal

sis of Idioms.” In: J. Puhvel (ed.),

ubstance and

tructure o

Lan

Los An

eles, Berkele

: Univ

rsit

of California Press.

elte

Werner. 1996.

nglische Morphologie und Wortbildung

edition. Frankfurt/M

Berlin

Bern

ew York, Paris, Wien: Peter Lan

, Alison. 2002. Formulaic Lan

e and the Lexicon. Cambrid

e: Cambrid

e Universit

Press.

tekauer

P. and R. Lieber

(

)

, Handbook of Word-Formation

375—42

2005 Sprin

er. Printed in the Netherlands

ENGLISH WORD-FORMATION PROCESSES

Observations, Issues, and Thoughts on Future Research

ROCHELLE LIEBER

1. INTROD

CTION

Previous chapters have looked at a wide variety of theoretical approaches to

word formation that have been developed over the last half century. The purpose of

his chapter is not primaril

theoretical. Rather here we turn the spotli

ht on En

lish

and look at the wa

s in which these theoretical developments have enriched our

nowled

e of word formation in En

lish. The focus of this chapter is descriptive,

but not descriptive in the sense that we merely review the data. Instead I explore the

ways in which theoretical debates have been played out on these data, and in which

he data have raised issues for various morphological frameworks. Where possible, I

also attempt to highlight areas in which f

rth

r r

ese

h n

eeds

obedo

ill

oncentrate here primarily on productive processes of word formation such as

ompounding, affixation and conversion,

eaving aside minor word formation

processes such as blendin

, clippin

, and back formation. M

focus will be

nchronic. If the reader wants an extensive diachronic description of En

lish word-

formation processes, Marchand (1969) is

ill th

ebes

t r

esou

aila

ynchronic account of English word formation processes can also be found in

Adams

(

2001

)

NDIN

ne of the principal and most heavily studied ways of adding words to the

English lexicon is compounding. Two forms of compounds are often distinguished:

ynthetic compounds (also called verbal, d

erbal, or verbal nexus compounds) are

nes in which the second stem is derived from a verb, and root compounds (also

alled primar

compounds) are ones in which the second stem is not deverbal.

Synthetic compounds are illustrated in

(

)

; root com

ounds in

(

)

(1) truck driver, gift-giving, wind blown, revenue enhancement, waste

disposal

(2) do

bowl, file cabinet, red hot, sk

blue, blackboard, bab

sit

Synthetic compounding is highly productive in English, as is the root

ompounding of nouns. Noun-ad

ective (

ky blue), ad

ective-noun (

blackboard

), and

ective-ad

ectiv

e (red hot

) root compounds are also relatively productive. Root

376

HELLE LIEBER

ompounds of other categories are harder to form and relatively unproductive (fo

exam

le, verb-verb com

ounds such as

tir-fr

or noun-verb com

ounds such as

abysi

.1 Determinin

what counts as a compound

ddl

, it is not eas

to come up with criteria that unequivocall

distin

uish all

and onl

compounds in En

lish, as has been pointed out b

Marchand (1969), Levi

(1978), Bauer (1978), Lieber (1992), and man

others. In fact, Bauer (1998)

oes so

far as to argue that there are no criteria,

ither individually or collectively, which

nequivocally serve to distinguish compounds from phrases. Among the criteria that

ave been proposed for compoundhood in En

ish are stress, spelling, lexicalized

eaning, unavailability of the first stem to syntactic processes such as inflection,

anaphora, and coordination, and inseparability of the first and second stems.

Spelling and lexicalization are universally agreed to be poor criteria fo

ompoundhood. En

lish spellin

is notorious

erratic in its treatment of

ompounds, with a sin

le compound someti

s bein

written as two separate

words, two words h

phenated, and one word (Bauer 1998: 96 cites the example of

aisy wheel

daisy-wheel

and

aisywheel

all of which are given as possible

spellings in the

xford Dictionary of New Words). Neither can compounds be

distinguished from phrases on the basis of lexicalization; while many compounds

ave lexicalized meanings (

blackboard

green hous

), compounding is highly

productive and new compounds are very oft

n compositional in meaning, at least

xt i

tak

n int

ccou

Stress often – but not alwa

s – distin

uishes compounds from phrases in

lish. Both root and s

nthetic compounds are t

picall

stressed on the first or left-

óg bowl

ápple cak

úck d

ive

móth eaten

. But not all compounds

rece

ft-han

ess.

Root compounds consisting of two ad

ectives (e.g.,

icy cold

lue-green

)

often do not, and instead seem to have level stress. Further, even some

noun-noun compounds do not exhibit left-hand stress: ápple cake does

but apple pí

s stressed on the right-hand stem. Bloomfield (1933: 180) goes so far as to insist

hat only combinations of two stems with left-hand stress are compounds, so that

ice

cream

is a com

ound when

ronounced as

íce c

eam

but a

hrase when

ronounced

ice c

éam.

Gie

erich (2004)

ives an excellent review of this controvers

and

proposes a somewhat less radical version of Bloomfield’s anal

sis.

Similarly, although the first stem of a compound is usually inflectionless, there

are examples like

children’s hour

girls’ clu

that clearly carry an inflection. The

first stem of a compound is also typically nonreferential; the first stem in

cat love

does not refer to any specific cat. Allen (1978: 113) claims that “individual elements

f compounds … generally cannot function i

dependently with respect to syntactic

processes.” But as Bauer (1998: 72) illustrat

, first stems can occasionally serve as

discourse antecedents for

ronouns

(

o I hear

ou’re a real cat-lover. How man

ou have now?

[

ame show, cited b

Ward et al. 1991: 47

])

. First stems of

arent com

ounds can also

ccasionally be conjoined,

specially, as Bauer points

ENGLISH WOR

FORMATION PROCESSE

ut (1998: 76) when they belong to the same semantic domain (for example,

cat and

og show, medical and life insuranc

nseparability is perhaps the strongest

est of compoundhood. The two stems of a

ompound are relativel

immune to separation b

a modifier:

truck

ast driver,

blue li

reen. This does not

of course

mean that the second stem of a

ompound cannot itself be complex; Bauer (1998: 73) cites examples like [Chines

[

jade figure

[[

]]. What it means is that an arbitrary modifier cannot be inserted between

elements of a com

ound and still maintain com

oundhood. So

[Chinese dirty

[

jade

[[

figur

]

] is unacce

table.

e acknowledge, however, that the line between compounds and syntactic

phrases in English is more labile that we

ight wish. In fact, one consequence of the

ack of unequivocal criteria for distinguishing compounds from phrases in English is

hat a controvers

has arisen over items like those in (3):

(

)

a who’s the boss wink

an if-you-only-want-to know sneer (From Ripley, Mike, Angels in

New York: St Martin’s Press 1991, 10,

uoted in Bauer and

enouf

(

2001: 108

))

number of theorists have considered forms like these to be ‘phrasal’

ompounds, that is, compounds in which the first element is a s

ntacticall

formed

phrase; see for example Botha (1980), Toman (1983), Hoeksema (1985, 1988),

Sproat (1985), Lieber (1992). This desi

nation would of course be problematic for

strictly lexicalist frameworks such as those of Selkirk (1982) and Di Sciullo and

illiams (1987). More recently, however, Bauer and Renouf refer to the second

form in

(

)

as an exam

le of a

hrase that functions as a

remodifier

(

see also

Spencer chapter, this volume). It is difficult to decide between the two alternatives,

owever. On the one hand, such forms are quite productive, occurring often on the

fly in spoken English, and more often tha

one might expect in written English. The

forms themselves are not lexicalized, nor is the initial phrase restricted to lexicalized

tems. On the other hand, it’s hard to see how we would identif

them positivel

ompounds. Stress is unhelpful here; even i

typical root compounds left hand stress

s not obligatory, and phrasal compounds

xhibit even more variability in stress

placement. We saw as well that lack of

inflection is not

a good indicator of

ompoundhood in typical cases. We are left with the criterion of inseparability. I

find (4b,d) far better than (4a,c), for example:

(4) a. *a who’s the boss filthy win

b. a filth

who’s the boss wink

c. *an if-

ou-onl

-want-to know insinuatin

snee

d. an insinuating if-you-only-want-to-know snee

f these judgments are sound, and if it is indeed impossible to insert a rando

odifier between the first (phrasal) part and the second stem, then perhaps it is

possible to class these forms as a type of compound in English.

378

HELLE LIEBER

2.2 Root compounding

n English, root compounds can be found consisting of combinations of the open

ate

ories N, V, and A; (5) (from Lieber 1992: 80)

ives examples:

(

)

N N file cabinet, towel rack, catfood, steelmill

N A sky blue, leaf green, stone cold, rock hard

A A icy cold, red hot, green-blue, wide awake

A N hard hat, bluebird, blackboard,

oorhouse

A V dry farm, wet sand, double coat, sweet talk

N V handmake, babysit, spoonfeed, machine wash

V N drawbridge, cutpurse, pickpocket, pull toy

V V stir-fr

, blow-dr

ump shoot,

ump start

f these, onl

the first four t

pes are productive, with N N compounds bein

far the most productive compound type in English. Compounds containing V as one

r both members are barely productive, although Bauer and Renouf (2001: 110) cite

neologisms such as

ry-burn

test

-r

elease

thumb

and

slow

bake

in th

orpus-based study of compounds.

oot compounds are syntactically, and usually semantically right-headed. This

eans that in general root compounds are endocentric, being interpreted as

pon

ms of their second stems. So a

file cabinet

denotes a kind of cabinet,

blu

s a sort of blue

and to dr

arm is a method of farmin

. But En

lish does have

exocentric root com

ounds: com

ounds like

red head

airhead

are not types of

ead, but are metonyms for types of people, in the former case a person with red

air, and in the latter case a forgetful, flighty person. This type of exocentric

ompound is traditionally referred to as a bahuvrihi compound

r an

(2001) also find ample neologisms of another sort of exocentric compound, that is,

lik

angel-beast

kick

which are used as premodifiers of nouns (fo

xample, ‘this unconscious an

el-beast division’ (2001: 113) or ‘a ver

kick-arse

attitude’

(

2001: 105

))

lish also has dvandva or copulative compounds. In the interpretation of this

ype of compound the stems are given equal weight. For example, a

roducer

director

is both a producer and a director, and a

fighter bomber

is equally a fighter

and a bomber. Note that it might be said that at least some exocentric and copulative

ompounds are endocentric in terms of their syntax, if not their semantics, as the

yntactic category of the compound still co

responds to the category of the second

tem. Such is the case with compounds like

redhead

producer-director

but

perhaps not with cases like

kick

angel-beast

which appear as premodifiers.

See Štekauer (1998) for an argument that there are i

fact no exocentric compounds in English. Within

is onomasiologocial theory of word formation, Šte

auer argues that these compounds are subject to a

process of ‘elliptical shortening’ which is

t in fa

t a matt

rmati

ENGLISH WOR

FORMATION PROCESSES

379

.3 S

nthetic compoundin

, like root compoundin

, is ri

ht headed, and is hi

productive in English. (6) gives some examples (taken from Lieber 1992: 81):

(

6) a. truck driver, meat eater, cat lover, flower grower, gift give

b. truck driving, meat eating, cat loving, flower growing, gift giving

c. power driven, moth eaten, well loved, home grown, law given

d. tax evasion, revenue enhancement, school closure, waste disposal

t is

enerall

reed that where the deverbalizin

suffix is -

-in

as in

(

4a,b), or one of the less productive nominalizers like -ion

-ment

-ure

as in

(

4d), the first stem is typically interpreted as an internal argument of the verb ste

(

so a

cat lover

is someone who loves cats). If the internal argument is manifested

yntactically, as in the home-growing of tomatoes

or if the verb stem lacks an

nternal argument, as in

ake-swimming

the first stem of the compound may be

nterpreted as an ad

unct. An ad

unct interpretation is also frequent when the suffix

s the passive

–en

as in (4c). It is generally said that the first stem in a synthetic

ompound cannot be interpreted as the sub

ect of the verb and therefore that

ompounds like

*girl-swimming

are unattested. While this seems to be true more

ften than not, Lieber (2004) notes examples like cit

emplo

in which a sub

ect

nter

retation is natural. Bauer and Renouf (2001) find several convincing

eologisms in their database, including

blood-pooling

(“The make-up man is surely

up for some technical prizes for the quality

f hi

and blood-pooling” (2001:

118

))

, and

director-buying

(“It was helped on Tuesday by director-buying and

ained a further 12.5p.” (2001: 119)). Jespersen (1965: 143) points out compounds

n which the second stem is a noun converted

(

or zero-derived

)

from a verb and in

which the first stem likewise bears a sub

ect relation to the second stem: sunrise

break, ni

all

At various times, theorists in the generative tradition have tried to capture the

elationship between the first and second stems in synthetic compounds in a numbe

f ways, including Roeper and Siegel’s (1978) First Sister Princi

Selkirk’s

(

1982

)

First Order Projection Hypothesis

and Lieber’s

(

1983

)

Argument Linkin

rincipl

, and through movement analyses such as those in Roeper (1988) and

Lieber (1992). Lieber (2004) gives a lexical

emantic account of the interpretation of

nthetic compounds (see section 1.4). We w

ll return to these theoretical positions

ow.

.4 Structure and inter

retation

he structure of root compounds is relatively uncontroversial: the righthand ste

s the head, and thus contributes category and morphosyntactic features to the

ompound as a whole.

380

HELLE LIEBER

lthou

h earl

enerative theorists attempted to derive root compounds

ransformationall

from complete sentences (Lees 1961, Kastovsk

1969, Brekle

1970) or to reduce their interpretation to a fixed number of t

pes (Levi 1978), in

ore recent years it has generally been a

reed

that th

fix

ed se

manti

elationshi

between the two stems of a root com

ound. In fact, there are two main

bservations we might make about the int

retation of root com

ounds: first, that

he lefthand stem receives a generic interpretation (that is, it is non-referential), and

econd, that the compound as a whole denotes a subset of the denotation of the

econd stem. For example, in the compound

catfood

fir

does

t r

r t

particular cat, and the compound as a whole denotes a sort of food, in this case,

food intended for cats. As for exocentric compounds like

airhead

, Booij (1992)

points out that morpholo

ical theor

need not sa

thin

special: whateve

principle of metonymy allows us to refer to players in a pick-up ball game on the

street as the ‘shirts’ (i.e., the guys in shirts as opposed to the ‘skins’, the ones

without shirts

)

will also allow us to account for the interpretation of bahuvrihi

ompounds.

Perhaps more problematic are cases like

kick

cited by Bauer and Renouf

(2001) as premodifiers; here meton

is not relevant, and the issue is whethe

kick

is to be treated as an ad

ective in a phrase like ‘a ver

kick-arse attitude’ o

as somethin

else. To date, I know of no particularl

satisf

anal

sis of forms o

sso

The structure and interpretation of synthetic compounds have been a matter of

some controversy among generative morphologists at least. Two basic structures

ave been proposed:

(

)

the deverbal noun

driver

is compounded with the noun

uck

whereas in

(8b) a compound is first formed of the noun

uck

and the verb

ive

h th

noun-formin

suffix -

is then attached. Lieber (1992: 85) discusses the relative

rit

f th

wo s

es:

The structure in [(8b)] has the advantage that it all

ows us

ay that an obligatory

omplement of a verb is satisfied in the first projection above the verb in synthetic

ompounds, just as it is in phrasal syntax. Such a structure has in fact been proposed by

(

)

N N A A

A A A

file cabinet hard hat sky blue blue green

(

)

a. N b. N

N N V

ve e

r tr

ENGLISH WOR

FORMATION PROCESSE

381

Lieber (1983). As has been pointed out by Booij (1988), however, this structural

analysis suffers from a rather serious drawback

in that it claims, at least implicitly, that

there should be verbal compounds like

uckd

ive

in En

lish. Yet this sort of compound

s not formed productively in English. The few verbal compounds that do exist (

abysit,

condition

and the like) are

enerall

acknowled

obe

esu

rmati

from s

nthetic compounds

(

sitter, airconditioner

etc.). In other words, a

ramma

hich

enerated [(8b)] as the structure for s

nthetic compounds would have to

vergenerate large numbers of nonoccurring (and impossible) verbal compounds.

The majority of generative morphologists, then, have chosen (8a) as the likely

tructure of synthetic compounds, among them Selkirk (1982), Booij (1988), and

Lieber

(

1992, 2004

)

Theorists’ choice of structure has to some extent depended upon the formal

echanism proposed to explain key observations concerning the interpretation of

nthetic compounds, namel

that the left-hand stem is t

picall

interpreted as the

ect of internal ar

ument of the verb (as in

asket makin

)

, unless the verb does

ot allow an internal ar

ument (for example, ocean swimmin

)

or the internal

argument is discharged syntactically (as in hand-making of baskets

)

. In the latte

ase

the first stem often bears an adjunct relationship to the verb.

n perhaps the earliest generative treatment of synthetic compounds, Lees (1961)

proposes a transformational derivation for synthetic compounds. Synthetic

ompounds arise from full sentence deep stru

tures and transformations successively

are down those sentential bases. A com

ound like

apple eater

derives from a deep

ese

the man eats a

les as do the s

nthetic compound apple eatin

and the root com

ound eatin

appl

. Marchand (1966) offers stron

contemporar

riticism of Lees’ analysis, pointing out

he arbitrariness of both the underlying

structures and the transformations proposed. His own semantically-based analysis,

ost fully worked out in his (1969) Categories and Types of Present-Day English

ord Formation

is in some way inspired by Lee

however

in that he classifies

synthetic compounds on the basis of grammatical relations that correspond to those

anifested in sentences (sub

ect, ob

ect, verb). Still, in the generative framework,

Lees’ work continued to be refined in the transformational anal

ses of compounds

proposed b

Levi (1978) and Bauer (1978) albeit in a far more principled and

onstrained wa

Subsequent generative analyses have

apitalized on theoretical changes in

generative syntax. With the advent of lexicalism in the mid-seventies, Roeper and

Siegel (1978) proposed a ‘lexical transformation’ which modifies the

subcategorization frames of deverbal nouns to result in synthetic compounds:

(

9) Roeper and Siegel (1978: 209)

[[empt

] + verb + affix]

[

+ word

]

W =>

[[

word

]

+ verb + affix

]

1 2

4 2

30 5

he lexical transformation in (9) takes a deverbal stem which has an empty slot

for a preceding stem and fills that slot w

th th

ewo

h f

ows

edeve

tem. This is intended to capture the generalization that the first stem of synthetic

where W ran

es over subcate

orization frames and

tan

r l

ate

ories N, A, Adv