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

Подождите немного. Документ загружается.

NDRE

PEN

Hockett

Charles F. 1958. A Course in Modern Lin

uistics

millan

Huddleston, Rodney and Pullum, Geoffrey K. 2002.

he Cambridge Grammar of the English Language

ambridge: Cambridge University Press.

Hudson, Richard. 1987. “Zwicky on heads.” Journal of Linguistics 23

109-32.

Hudson, Richard. 2003. “Gerunds without phrase structure.”

Natural Language and Linguistic Theory

579-615.

ackendoff, Ray, S. 1997. “Twistin’ the night away.” Languag

534-559.

ackendoff, Ray, S. 2002a. “English p

ticle constructions, the lexicon, and the autonomy of syntax.” In:

ehé et al.

(

eds.

)

, 67-94.

ackendoff, Ray, S. 2002b. Foundations of Languag

xford:

xford

versity Press.

ne, Richard. 1994.

he Antis

mmetr

nta

Cambrid

e, MA: MIT Press.

Kiparsk

, Paul. 1997. “Remarks on Denominal Verbs.” In: A. Alsina et al (eds.), 473-500.

Kopt

evska

a-Tamm, Maria. 1993.

ominalizations. London: Routled

Koptjevskaja-Tamm, Maria. 2003. “Action nominal c

nstructions in the languages of Europe.” In: F.

Plank

(

ed.

)

, Noun Phrase Structure in the Languages of Europe

Berlin: Mouton de Gruyter, 723-

59.

Lapointe, Steven, G. 1993. “Dual lexical categories

nd the syntax of mixed category phrases.” In: A.

Kathol and M. Bernstein

(

eds.

)

, Proceedings of the East Coast States Conference on Linguistics

1993, 199-210.

Levin, Beth and Rappaport, Malka. 1988. “Non-event -

nominals: a probe into argument structure.”

Linguistics

26, 106

Lewis, Geoffrey L. 1967.

urkish

rammar.

xford:

larendon Press.

Lieber, Rochelle. 1983. “Ar

ument linkin

and compounds in En

lish.” Lin

uistic Inquir

251-286.

Lieber, Rochelle. 1988. “Phrasal Compounds in E

lish and the Morpholo

-S

ntax Interface.” In:

Proceedings of the Twenty-fourth Meeting of

the Chica

o Lin

uistic Societ

, Parasession on

reement in Grammatical Theor

202-220.

Lieber

Rochelle. 1992. Deconstructing Morphology

Chicago: University of Chicago Press.

Malouf, Robert. 2000a. “Verbal gerunds as mixed cat

ories in Head-Driven Phrase Structure Grammar.”

In: R. D. Borsley (ed.),

The Nature and Function of

Syntactic Categories

yntax and Semantics vol.

2. San Diego: Academic Press, 133-166.

Malouf

Robert. 2000b.

ixed Categories in the Hierarchical Lexicon. Stanford University: Center fo

he Study of Language and Information.

Marchand

Hans. 1969.

he Cate

ories and T

pes o

Present-Da

lish Word-Formation. A

Synchronic-Diachronic Approach

Munich:

. H. Beck.

Pollard, Carl and Sag, Ivan. 1994.

ead-Driven Phrase

tructure

rammar. Chicago: Chicago University

ess.

Pollock

Jean-Yves. 1989. “Verb movement

Universal Grammar

and the structure of IP.” Lin

uisti

nquir

365-424.

Pullum, Geoffrey K. 1991. “English nominal gerund phrases as noun phrases with verb-phrase heads.”

Linguistics 29

763-799.

adford

Andrew. 1997.

yntactic Theory and the Structure of English

Cambridge: Cambridge

niversity Press.

amchand

Gillian and Svenonius

Pet

. 2001. “The Lexical Syntax and Lexical Semantics of the Verb-

Particle Construction.” Proceedings of the 21 West Coast Conference on Formal Linguistics

Mikkelsen and C. Potts

(

eds.

)

, So

erville

MA: Cascad

lla Pr

,38

7–

400

Randall

Janet. 1982.

orphological structure and language acquisition

Unpublished PhD dissertation,

niversity of Massachusetts, Amherst. [Published by Garland Publishers, New York, 1989]

Rappaport Hovav, Malka and Levin, Beth. 1992. “-er no

nals: Implications for the theor

of ar

ument

tructure.” In: T. Stowell and E. Wehrli

(

eds.

)

ntax and Semantics. V

. 26: S

ntax and the

exicon

San Die

o: Academic Press, 27-153.

Roeper, Thomas. 1987. “Implicit ar

uments and the head-complement relation.” Lin

uistic Inquir

18,

267-310.

Roeper, Thomas. 1993. “Explicit syntax in the lexicon

The re

resentation of nominalizations.” In: J.

Puste

ovsk

(ed.),

emantics and the lexicon

Dordrecht: Kluwer Academic Publishers

185-220.

Roeper, Thomas and Siegel, M.E.A. 1978. “A lexical transformation for verbal compounds.” Linguisti

nquir

9, 199-260.

Scalise, Sergio. 1984.

orphology in Generative Grammar

ORD

FORMATION AND SYNTA

Selkirk

Elizabeth O. 1982.

he Syntax of Words

Cambridge, MA: MIT Press.

Spencer, Andrew. 1999. “Transpositions and argument structure.” In: G.

ooij and J. van Marle (eds.),

Yearbook of Morphology 199

Dordrecht: Kluwer Academic Publishers

73-102.

Spencer, Andrew. 2000. “Morphology and Syntax.” (Art. 34) In: G. Booij, Ch. Lehmann and J. Mugdan

(

eds.

)

orpholo

ie. Ein internationales Handbuch

ur Flexion und Wortbildin

. 1. Halbband.

Morphology. An International Handbook

on Inflection and Word-Formation

olume 1

rlin

alter de Gru

ter, 312-335.

Spencer, Andrew and Zwicky, Arnold (eds.). 1998.

andbook of Morpholog

Oxford: Blackwell

roat, Richard. 1985.

n deriving the lexicon

ublished PhD dissertation, Massachussetts Institute of

Technology.

Štekauer

Pavol. 1998. An Onomasiological Theory of English Word-formation

Amsterdam: John

enjamins.

Strauss

Steven L. 1982. Lexicalist Phonology of English and German

ati

oivonen, Ida. 2002. “Swedish particles and syntac

ic pro

ection.” In: Dehé et al. (eds), 191-210.

oivonen

Ida. 2003. Non-Projecting Words. A Case Study of Swedish Particles

uwe

oman

Jind

ch. 1983. Worts

ntax: Eine Diskussion aus

ewählter Probleme deutscher Wortbildun

Tübin

en: Nieme

er Verla

oman

Jind

ch. 1998. “Word syntax.” In: A. Spencer and A. Zwicky (eds.), 306-321.

an Hout, Angeliek and Roeper, Thomas. 1998. “Ev

ts and As

ectual Structure in Derivational

Morphology.” In: H. Harley (ed.), Papers from the Upenn/MIT Roundtable on Argument Structure

and Aspect,

ol. 32. MIT Papers in Linguistics. Cambridge, Mass, 175-200.

älchli

Bernhard. 2003.

o-Compounds and Natural Coordination. Unpublished PhD dissertation,

niversity of Stockholm [revised version forthcoming from Oxford University Press].

estcoat, Michael. 1994. Phrase structure, lexical

haring, partial ordering and the English gerund.

Proceedings of the Berkeley Linguistics Societ

20,

7-5

oon, James. 1996. “Nominal gerund phrases

n English as phrasal zero derivation.” Linguistics

34, 329

wick

, Arnold. 1985. “Heads.”

ournal o

Lin

uistics 21

1-29.

tekauer, P. and R. Lieber (eds.), Handbook o

Word-Formation,

—124

2005 Springer. Printed in the Netherlands

HANS MARCHAND AND THE MARCHANDEANS

DIETER KA

OVS

1. INTROD

CTION

Hans Marchand’s contribution to the theor

of word-formation in general and to

the description of English word-formati

n in particular has unquestionably been

extremely influential, and his handbook

he categories and types of Present-day

nglish word-formation

(

1st ed. 1960, 2nd. ed. 1969) is still an unsurpassed

andmark in the field. A discussion of the basic assumptions underlying his approach

s therefore certainly appropriate in the context of this volume. But what about the

term ‘Marchandean’, which was suggested by the editors of this volume? What does

t m

an t

obe

a ‘Mar

han

an’? This certainly needs some kind of specification. Is it

the fact that one has worked directly under Marchand’s supervision as a research

assistant, like Herbert Ernst Brekle, Leonhard Lipka, myself and Gabriele Stein (the

ames are given in chronological order of appointment)? Does it mean that someone

as been very much influenced by his ideas, like Klaus Hansen, although he neve

et him personally? Or does it mean that someone has worked as part of a circle of

uists (nowada

s sometimes called the ‘Tübin

er Schule’), to which Marchand

also belon

ed, but where other influences (e.

. Mario Wandruszka, Eu

enio Coseriu

n Romance and

eneral lin

uistics, Ha

s-Jür

en Herin

er, Otmar Werner in

German linguistics) had also been very strong, as in the case of Hans Martin Gauger,

Franz Hundsnurscher

Wilfried Kürschner or Christian Rohrer? And even in the case

f Marchand’s assistants, these latter linguists and other influences (especially fro

enerative-transformational grammar) are clearly visible, since none of them would

take over Marchand’s approach complet

y unchanged. Tübingen in the 1960s and

1970s was a hotbed of modern lin

uistics, with an active lin

uistic circle and a lot of

ross-fertilisation also due to

uests from outside. Therefore, the ter

‘Marchandeans’ is

erha

s somewhat

roblematic. On the other hand

there is no

denying that Marchand did have a great influence on many of us working on word-

formation in Tübingen in the 1960s and 1970s and also later on, and therefore the

term ‘Marchandean’ has a certain amount of

ustification, especially in view of the

fact that there is a certain common theoretical basis underlying the work done by us.

In the present context, it is of course not possible to take into consideration the more

eneral influence of Marchand on many co

leagues in Tübingen, who had not been

embers of the En

lish Department, and I can onl

deal with the ‘inner circle’, i.e.

Mar

han

’

ese

h a

tant

swe

ll a

Klaus Hansen, who Marchand re

arded as

is oldest pupil, althou

h Hansen knew M

chand onl

readin

his publications

and by an intensive correspondence, and, second-hand, through his contacts with

rekle, Lipka, Stein and myself. Moreover, I will concentrate on contribution to

word-formation and not to the many other fields in which they have been active.

IETER KA

OVS

lso

I will focus on those areas in which the Marchandeans have added to o

odified Marchand’s theory rather than sk

tching their overall theoretical approach,

which has already been done in Štekauer (2000).

2. HAN

MAR

HAND

.1 Theoretical

ramewor

Hans Marchand’s theoretical framework

which he had develo

ed in the 1940s,

ad its primar

roots in European

uctu

alism

but was also influenced b

merican structuralism: as a student of Leo S

itzer in the 1930s he had come into

ontact with European/French structuralism, but during his stay in the U.S. in the

ate 1940s and 1950s he had become familia

ith th

f Am

tructuralism, whose basically anti-semantic bias he ob

ected to quite vehemently.

This basically descriptive-structuralist approach to word-formation, which

haracterises the first edition of his handbook

was later-on modified to a certain

extent b

the attempt to inte

rate certain insi

hts of

enerative-trans

ormational

grammar

especiall

in connection with a critical discussion of Lees’s (1960)

pro

rammatic boo

The

rammar o

lish nominalizations

n Marchand

(

1965a,

b; Lees 1966), which eventually found its way into the second edition of his

andbook

(

Marchand 1969

)

.2 Synchronic approach

Marchand’s word-formation theory is

ynchroni

(

in contradistinction to

previous approaches, which had all been diachronic, cf. Jespersen 1942, Koziol

1937), althou

h he deals with

histo

ical

aspects in his handbook as well, cf. its

sub

titl

A synchronic-diachronic approach, and it is based on the following basic

assum

tions:

1) S

nchronic description has priorit

; onl

after a s

nchronic description has been

provided can one look at the history of the patterns characterising a synchronic

ystem.

2) A synchronic description of

ord-formation patterns

obeb

sed o

n th

otions of motivation/analysability, patt

n, and productivity, which involves as

iti

nal

ntral n

n that

f th

agma

a combination of a

dete

minant

(

modifier

)

and a

dete

minatum

(

head

)

3) A s

nchronic description of word-formation has to reco

nise the existence of

stematic morphophonemic alternations

which have to be treated in terms of

their s

nchronic function, not their historical ori

in. This mi

ht require the

For a more detailed account of its genesis,

f. Kastovsky (1999). For a critical, comprehensive

evaluation, cf. Štekauer

(

2000: 29-48

)

ANS MARCHAND AND THE MARCHANDEANS

101

distinction between two levels of word-formation

a native and a non-native one

hough not in a purely etymological, but in a structural-functional sense.

2.3

otivation

The most central concept in Marchand’

word-formation theor

is that of

motivation

which goes back to Saussure and was further elaborated by Bally (1944).

This concept, nowadays also referred to as the compositionality principl

assum

hat simple linguistic items (signs, morphemes) are in principle

arbitrary/unmotivated with regard to the relationship between form and meaning

(with the exception of onomatopoeia), while complex linguistic constructions (at

whichever level) are in principle

relatively motivated

because they can be

nterpreted semanticall

on the basis of the knowled

e of the meanin

s of thei

onstituents and some

eneral underl

pattern.

This assumption has the followin

consequences:

1) Meanin

is as important as form, which is wh

the simple lin

uistic si

n, the

orpheme, is re

arded as a combination of for

and

meaning, i.e. as a “two

facet si

n, which means that it must be based on the si

nifi

ate/si

nificant [...]

elationship posited by Saussure” (Marchand 1960: 1). This point of view

differs fundamentally from the mainstream American structuralism of the

1940s and 1950s, which was basically form-oriented and treated morphemes as

purely distributionally defined formal entities

(

Kastovsky 1997b). Thus

Marchand re

ected the then cu

rent analysis of verbs like

receive

deceive

conceive

;

etain

detain

contain

etc. as bi-morphemic (cf. e.

. Harris 1942:

1), because the alle

ed ‘morphemes’ did not have an

identifiable meanin

and therefore could not be re

arded

uistic si

ns (cf., e.

., Marchand 1955

[

1974]: 180f.

2) Only morphologically

and

semantically motivated combinations can give rise

o new morphologically and semantically analysable formations, i.e. word-

formation is based on formal and semantic analogy, which is equivalent to

aying that there has to be an underlying morphosemantic pattern

which is – at

east to a certain extent – also productive in so far as it allows the creation of

ew formations. Consequentl

, word-formation should onl

deal with

nchronicall

productive patterns

The centrality of the concept of motivation also plays an important role in

Marchand’s description of onomatopoeia. In contradistinction to Saussure

who had

ued that onomatopoeia, bein

simple si

ns, were unmotivated, Marchand ar

ues

for a considerable amount of motivation in this domain. The importance which this

tance had for him is documented by the fact that one of his first theory-oriented

word-formation

ers

(

Marchand 1949

)

deals with onomato

oeia. This is where the

notion of productivity is explicitly mention

r th

fir

t tim

dbe

that a

imilar f

rm-

ed analysis was still advocated by Halle (1973: 10), who suggests

mentations such as serendip + i + t

, vac + ant, tot + al, bro + ther, be + liev

and cf. Lipka’s

(

1975a, b)

ustified critic

sm of this approach.

2 DIETER KA

OVS

hat fr

m th

ound symbolism remained one of his major interests, cf. the

ective cha

ters in Marchand

(

1960, 1969

)

.4 Morphonological alternations

The notion of productivity is also linked with the role of

orpho(pho)nological

(morphophonemic)

alte

nations

in word-formation

which Marchand addresses fo

he first time in Marchand (1951), ar

uin

that “an opposition which does not form a

orpholo

ic pattern, is not relevant to morphonolo

, but to phonolo

onl

”

(Marchand 1951: 88). Thus, only productive, pattern-forming alternations (whethe

n inflection or derivation) should be dealt with in morphonology, whereas

al in

tan

ces suc

h a

beau

belle

hould be treated as straightforward

allomorphy (as we would call it today). Thi

that i

tan

ces suc

h a

deceive

eception, resume : resumption should be regarded as purely lexical (allomorphic)

alternations, which are historically due to independent borrowings and are therefore

neither part of word-formation, nor do the

represent

enuine s

nchronic

orphophonemic alternations.

Marchand furthermore ar

ues that this phenomenon reflects the fact that “most

European languages [...] have two for

ative

rinci

les”

(

Marchand 1951: 92

)

, a

native and a non-native one. The crucial factor here is not etymology, i.e. whether

he respective item or pattern is borrowed, but whether a given formation is

synchronically formed according to produc

ive native patterns o

according to the

patterns of the source language, e.g. Latin and/or Greek. Thus

cultivate

cultivatable

educate : educatable are based on a native pattern, whereas

avi

ate :

avi

able, communicate : communicabl

are based on a non-native pattern.

.5 The concept of syntagma

The conce

t of motivation has an add

tional dimension. Marchand had ado

ted

his concept from Bally (1944), who had linked it to the notion of

agma

i.e. a

sign combination based on a determinant/determinatum structure. Consequently,

ord-formation is also based on this principle, i.e. the results of word-forming

processes are basicall

binar

and the

alwa

s have a head (determinatum). This is

also the basis for interpretin

so-called conversions of the t

walk

rid

clean

b as bi-morphemic, containin

zero-

hem

as head, which has repeatedly bee

rejected, cf., e.g., Pennanen (1971),

Lieber

(

1981: 126ff., cf. also Štekaue

2000: 246f.), or more recently Štekaue

(1996), but which is generally accepted by

he Marchandeans, cf., e.g., Kastovsky

(

1968, 1986, 1996

headedness

h lat

r intr

oduced

int

enerative word-formation by Williams (1981

)

, but it had already been present in

In the latter two

ers I have tried to show that there are also historical reasons for assuming a zero

orpheme/allomorph, because at an earlier sta

f the lan

e there e

sted overt morpholo

ical

aterial, viz. stem formatives, expressin

the r

pective function, which was lost due to phonolo

ical

processes, but leaving behind its f

unctional load: in other words, an

overt morpheme was replaced by

othin

, i.e. zero.

ANS MARCHAND AND THE MARCHANDEANS

103

merican structuralism

cf. the distin

wee

doce

ntri

cco

onsisting of a head and a modifier, and exocentric constructions

which did not

ontain a head. The earlier version of generative grammar with its phrase structure

ules did not allow for the specification of a head of a construction, and it was onl

within the framework of X-bar syntax that

headedness could also be characterised

formall

in the

enerative framework.

further conse

uence of the headedness

rinci

le and the consideration of the

ypological properties of a language with regard to the sequence of modifier and

ead in morphological syntagmas led Marchand to postulate that in the Germanic

anguages word-formation syntagmas were generally characterised by the sequence

f modifier/head, in contradistinction to the Romance languages, where also the

sequence head/modifier occurs. This means

hat prefixes can only act as modifiers

and not as heads

i.e. in formations such as

rost, discoura

e, enca

e, unsaddl

despite the chan

e of word-class involved, the prefix cannot act as head.

This

henomenon is also linked with Ma

han

’

nal

tin

wee

ansion an

ivation

(

Marchand 1967

)

, where derivation is treated as a

subcategory of trans

osition, again taken over from Bally (1944), and which cuts

across the purely formal distinction between compounding and affixation

(prefixation and suffixation). The latter is based on the formal status of the

onstituents of the word-formation syntagma (compounds consist of lexemes,

affixati

sco

f a l

bou

heme

(

affix

))

. The former is based

n the functional status of the head: in ex

ansions, the head can stand for the whole

ombination accordin

to the formula

B = B an

d cove

both compoundin

and

refixation, since in both cases the deter

nant just modifies the determinatum: a

steamboat

boat, precook

is a kind of

cook

With derivations

(

= suffixations

)

, on

he other hand, this formula is not a

licable, i.e. AB



B, and the suffix trans

oses

he base into a different lexical category, i.e. a

t an -

ally had generalised the syntagma principle to all linguistic levels, including the

leve

f th

ese

ce.

ssu

that th

ect functions as determinatum (head),

and the predicate as determinant (modifier). Moreover, he linked these functions to

he or

anisation of the information structure in a sentence: the topic (‘thème’) is

equivalent to the determinatum, and the comment (‘propos’) corresponds to the

determinant

and in the unmarked case th

is co

des w

ith the subject/predicate

dichotomy, cf.: “la phrase est un syntagme, de même que tout groupe de signes plus

rand ou plus petit, susceptible d’être ram

é à la forme de la phrase” (Bally 1944:

102). This is the original source of the notion of topicalisation

also discussed in

Marchand (1965a), although without using this term, which plays an important role

n Brekle’s, Lipka’s and my own work later on.

This interpretation opens the wa

for a s

ntactic interpretation of complex

exical items, a central issue in

enerative

ransformational

rammar (cf., e.

., Lees

For a more detailed discussion of the problems

nvolved in this anal

sis, cf. Marchand (1969: 134-

136), Hansen (1980), Kastovsky (1986), where the deri

ation is associated with a zero-determinatum

which allows the prefix to act as a determinant. The prefix as head is suggested, among others, in

Lieber

(

1981

)

and Williams

(

1981

)

which

oes a

ainst the

eneral t

polo

ical structure of the

ermanic languages.

104

DIETER KA

OVS

1960), which is already also suggested at

east implicitly in Marchand (1960), cf. his

emarks on the compound type house-keeping: “Cpds are chiefly based on a

‘predicate/object’ relation, but as such cbs are, on

rinci

le, nominalized sentences,

ther relations also occur”

(

Marchand 1960: 29

)

This syntactic dimension is also addressed in some articles on

ode

ivation

Marchand (1963a, 1963b), where he provides the following analysis of zero-derived

ouns: “Other sense

roups occurrin

are: 2) the idea of ‘ob

ect’, affected o

effected, personal or impersonal = ‘one who, that which is

ed’; 3

)

the idea of

‘a

ent’, personal or impersonal, material or immaterial = ‘one who, that which

’

;

the conce

t ‘adverbial com

lement’ = ‘

ce o

r in

ed w

ith th

erbal process’” (Marchand 1963a: 185). And of denominal verbs he says:

“

Denominale Verben bezeichnen

rinzi

iell Handlungen, in denen das zugrunde

iegende Substantiv syntaktische Funktionen

at. [...] das Substantiv fungiert als

sub

ektbezogenes Prädikatsnomen [...], als adverbiale Ergänzung [...], als Ob

ekt.”

(

Marchand 1964b: 105

)

. Marchand had thus realised that word-formation had a

ntactic dimension on the basis of Ball

’s s

nta

ma principle, before he was

onfronted with the

enerative-transformational approach to this domain advocated

n Lees

(

1960

)

.6 Generative-transformational influence

n the early 1960s, Herbert E. Brekle and Christian Rohrer caught on to this new

framework and got Marchand interested in it, and he began to explicitly integrate the

ntactic dimension into his framework, cf. Marchand

(

1965a, b

)

and Lees

(

1966

)

nfortunatel

, this controvers

was frau

ht with man

misunderstandin

Marchand did not reall

adopt the then prevalent

enerative-transformational

formalism, according to which word-for

ations were explicitly derived by

ransformations from underlying sentences, even though he used terms such as

transformation

underlying sentenc

This is why his terminology is somewhat

ambivalent, on the one hand speaking of ‘compounds reduced from sentences’ (the

enerative-transformational perspective) and ‘compounds reducible to sentences’

(the Bally perspective echoing the phrase “su

eptible d’être ramené à la forme de la

phrase”).

This s

ntactic perspective was then inte

rated – at least partl

– into the second,

evised edition of his book, whose su

sion was m

first

ob as Marchand’s

assistant in 1967. It has affected the cha

ters on com

ounds and on zero derivation,

but less so the

refixation and suffixation cha

ters. It can best be illustrated on the

basis of his approach to the analysis of compounds, introducing various levels of

analysis (Marchand 1969: 53) as well as the notion of types of referenc

(

Marchand

I m

ess

that at that tim

I al

understood some of the assumptions of

enerative-

ransformational grammar, taking the transformation

processes to operate on actual sentences rather

han on abstract underlying syntactic stru

tures, cf. Kastovsky (1968) and below.

For a detailed analysis of these misunderstanding

and Marchand’s equivocal use of terminology, cf.

Kürschner

(

1977

)

HANS MARCHAND AND THE MARCHANDEANS

105

1969: 32), also figuring prominently in the approaches of some of the Marchandeans

.7 Anal

sis o

compounds

Marchand assumes the followin

levels of anal

sis for compounds: morpholo

hape, morphologic structure, content at the level of grammatical deep structure,

ype of reference, content at the morphological level.

orphologic shape

deals with the morphological stat

us of the constituents of a

ompound, i.e. whether these are free (words) or bound (affixes or stems).

orphologic structure specifies the sequence of determinant and determinatu

for the combination in question. It should be added that is also describes the

mmediate Constituent structure at the mo

phological level, i.e. it would state that

lette

ite

as the morpholo

ical structure

lette

(

dm) as a

ainst

back

bench

(

)

onten

at the level of grammatical deep structure an

types of reference ar

nterconnected and are related to “an underlying sentence whose syntactic relations

hey mirror. This underlying grammatical

elationship will be called grammatical

eep structure” (Marchand 1969: 55). This underlying structure basically

orresponds to the notion of

nel sentence

n Chomsky (1957), i.e. it represents a

simple active sentence, and its syntactic st

ructure forms the basis of the semantic

nterpretation of the compound. This is most obvious with the so-called

bal nexus

compounds

which contain an overt v

bal element

such as

etter-writer

cock

fighting, draw-bridge, closing-time,

minal an

affixal

easily be associated with the syntactic functions they would have in an underlying

entence. But Marchand extends this also to purely nominal (i.e. non-verbal nexus)

ompounds like

oil well

steam boat

, etc., where an underlying verb has to be

econstructed in order to assign a syntactic function to the nouns

based on the

yntactic

alenc

of this underlying verb.

t this sta

e, the notions of topicalisation and t

pes of reference come into pla

ne s

ntactic part of the underl

sentence, the sub

ect, ob

ect, adverbial

omplement or predicate can be made the topic of the sentence, which will then

‘

rfa

’ a

ede

rminat

m an

dco

tit

the type of reference of the respective

ombination. Thus, an underlying structure

someone

(

)

eats

(

)

les (O) may

yield the following compounds: a

le-eat-er

(

= Subject Type), eating apple

(

Object

Type),

apple-eating

(Predication Type)

riting table

omeone writes somethin

at this table

(Adverbial Complement Type). Not

that Mar

han

purely syntactic functions, i.e. to what came to be called Argument Structure

n m

The term Predication Type is somewhat problematic, since it does not really refer to a nominal

syntactic function, as is the case with the othe

ypes. I had therefore already in Kastovsky (1968:

27f.) suggested to split up the predicate into a g

eral category of state, action or activity, which

would then surface as determinatum

and the r

aining verbal nucleus, which would act as

rminant

Thi

lat

rat

in Kastovsky (1976), and in Kastovsky (1982: 188f.),

adopting the analysis of complement sentences by

the Kiparsky’s (1970), where elements such as

FACT, ACT, ACTION, STATE, acting as head

ouns of the respective complement sentences,

wou

dse

f th

esu

rfa

ce de

rminata

IETER KA

OVS

ecent generative grammar, but he also in

udes ce

rtain a

iti

nal

manti

differentiations such as affected and effected object, adverbials of instrument, place,

ime, which corres

ond to Fillmore’s

(

1968

)

-structure cases an

eferred to as ‘thematic structure’

,‘

roles’ or ‘thematic grid’,

adopted by some Marchandeans later on, e.g. Kastovsky (1973, 1974, 1982).

In many instances, especially with verbal nexus combinations containing an

xplicit verb, this reference to the s

ntactic underl

structure at the same time

rovides a sufficient

semantic

anal

sis

lette

iter

is ‘someone who writes a

etter/letters’

mincemeat

s ‘meat which has been minced’

reezin

poin

s ‘

oint

when freezing occurs’. But in many othe

nstances

additional semantic information

s needed for an a

riate inter

retation of the formation in

uestion. This is

where content at the morphological level comes in. Thus, in the case of

raw-bridg

there is an underlying predicate-ob

ect relation, but it has to be added that this bridge

also incorporates the notion of ‘purpose’, viz. ‘bridge constructed to be drawn’; a

baker

may be someone who has just baked a cake, but may also refer to a

rofession, etc. Or, in the case of call-bo

he ori

inal interpretation was that of

eferrin

to a person callin

actors onto the sta

e (a profession) based on a sub

ect

–

predicate relation, which now – in analo

callgirl

– has been extended to a

predicate – object relation. These semantic modifications, which can be very

ecific, came to be treated under the rubric of ‘lexicalisation

institutionalisation

diomatisation’ by Lipka and myself, cf. below, and are necessary in order to

account for the lexical properties of many of these formations which go beyond the

underlying syntactic relations.

.8 Precursor of Lexicalist Hypothesis

s beco

e obv

ous o

f thi

rat

er sketchy account, for Marchand the

yntactic analysis of word-formation syntagmas was not equivalent to

transformational derivation

despite the terminology used by him. Rather, this kind

f analysis was regarded as a heuristic principle in order to make the relationship

between the morphological and the syntactic level explicit and to provide a bette

basis for the semantic interpretation of composites. Looked at by hindsight today,

Marchand’s revised word-formation theor

us mi

ht be interpreted as a precursor

hat i

ow c

all

lexicalist

word-formation, which also correlates word-

formation constituents with s

ntactic function, but without postulatin

transformational relationship between the syntactic and the lexical level.

Note that in this respect there is an ongoing cont

oversy as to whether argument structure or thematic

oles should be involved in the relationshi

ween word-formation syntagmas and corresponding

yntactic structures (cf. Kastovsky 1995).