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

THE LEXICALIST APPROACH TO WORD

-

FORMATION

18

7

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

Morpholo

g

ie / Morpholo

gy

: an international ha

n

dbook on inflection and word-formation

,

vol. 1.

B

erlin – New York: Walter De Gru

y

ter, 312-334.

Sproat, Richard W. 1985. On deriving the lexicon

.

Ph.D. dissertation. Cambridge (Mass.), MIT.

Štekauer, Pavol. 2001. “Beheadin

g

the Word? Please, Stop the Execution.” Folia Lin

g

uistica 34/3-4

,

333-

355

.

Stump, Gre

g

or

y

T. 2001. In

f

lectional morpholo

gy

.

Cambrid

g

e: Cambrid

g

e Universit

y

Press.

V

an Kemenade

,

Ans. 1987.

S

yntactic case and morphological case in the history of English

.

D

o

r

d

r

ec

ht

:

Fo

ri

s.

W

asow

,

Thomas. 1977. “Transformations and the Lexicon.” In: P. Culicover

,

Th. Wasow and A.

A

kmajian (eds), Formal Synta

x

. Academic Press 327-360.

W

illiams

,

Edwin. 1981. “On the notions ‘Lexi

c

all

y

Related’ and ‘Head of a Word’.” Lin

g

uistic Inquir

y

12, 24

5–7

4

.

Z

wick

y

, Arnold M. 1985. “Clitics and particles.” Lan

g

ua

g

e

61, 283

–

30

5.

Z

wicky, Arnold M. 1986. “The general case

:

Ba

s

i

c

f

o

rm

vs. de

fa

u

lt f

o

rm

.

” In

:

Berkeley Linguistic Societ

y

12

,

305–14.

Z

wick

y

, Arnold M. 1990. “Some Choices in the Theor

y

of Morpholo

gy

.”

I

n

:

Formal Grammar: Theor

y

and Implementation.

V

ancouver Studies in Cognitive Science

.

Vancouver: University of British

C

olumbia.

18

9

Š

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

f

Word-Formation, 18

9

—205

.

©

2005 Springer. Printed in the Netherlands

.

LEXEME-MORPHEME BASE MORPHOLOGY

RO

BERT BEARD AND MARK

VO

LPE

1. INTR

O

D

UC

TI

O

N

L

EXEME

-M

O

RPHEME

B

AS

E

M

O

RPH

O

L

OG

Y (LMBM) is a theory of morphology

w

hich claims that lexical morphemes, called L

exemes

,

and grammatical morphemes,

M

orphemes

,

are radically different linguistic phenomena. This hypothesis is based

o

n the properties distinguishing Lexemes and Morphemes listed in Table 1:

L

exemes Morphemes

B

elong to an open class

B

elong to a closed class

Ha

ve

r

e

al

wo

rl

d

r

e

f

e

r

e

n

ces

R

efer only to grammatical categories

Must be phonemically expressed

May be phonemically expressed

T

a

b

l

e1

Lexemes and Mor

p

hemes

The definitions of the two categories are simple: Lexemes are noun, verb, and

adjective stems. These items in all languages are manifested without exception as

sound-meaning pairings that refer to something in the real world. Any othe

r

m

eaningful linguistic phenomenon is a Morpheme and hence must refer to a

g

rammatical cate

g

or

y

; it cannot be used in reference to an

y

thin

g

in the extra-

l

in

g

uistic world. Morphemes refer exclusivel

y

to universall

y

available closed-class

g

rammatical cate

g

ories like Tense, Aspect, and Number and ma

y

consist of

i

ndependent phonemic st

r

ings (usually unaccented), affixes, infixes, changes in

accent or tone, or even predictable omissions (zero morphemes).

a

1

2. THE THREE BASIC HYPOTHESES OF LMBM

LMBM comprises three basic hypotheses:

T

he Separation Hypothesis

c

laim

s

that l

e

xi

c

al an

d

i

nfl

ec

ti

o

nal

de

ri

v

ati

o

n ar

e

processes distinct from phonological realization (affixation, etc.);

T

he Unitary Grammatical Function Hypothesis

c

laim

s

that th

e

r

e

ar

e

44

u

niversally available grammatical function

s

used

f

o

r

bo

th infl

ec

ti

o

nal an

d

l

e

xi

c

al

d

erivations

;

1

Since morphemes are often the result of reducin

g

a lexeme, lexemes in transition, servin

g

both as

l

exemes and morphemes, are not uncommon. In US English, for example,

have

i

s

a l

e

x

e

m

es

in

ce

it

does not behave like a morphemic auxiliary. In Br

i

tain

,

however

,

this verb behaves more like an

auxiliar

y

, which LMBM treats as a

morpheme: it may be contracted (

I've a new book

(

) and it may be

k

r

aised in questions

(

Have you a new book

).

k

1

90

R

O

BERT BEARD

&

MARK

VO

LP

E

T

he Base Rule Hypothesis

c

laim

s

that th

eu

ni

ve

r

s

al

c

ategories of word and clause

s

tructure must originate in a base component in order to explain both lexical and

s

yntactic (= inflectional) derivation most economically;

L

et us examine each of these hypotheses one by one.

2.1 The Separation H

y

pothesis

For centuries lin

g

uists stru

gg

led with a set of morpholo

g

ical eni

g

mas: (1) zero

m

orphemes, i.e., morphemes without phonolo

g

ical realization, as in the noun

a cook

de

ri

ved

fr

o

m th

eve

r

b

to cook

,

(2) empty morphemes, morphemes with no semantic

r

ealization, e.g.,

-al

in

l

s

yntact-ic-al

,

and (3) morphological asymmetry, the fact that

a single morpheme could have several functions, as the

-ing

in the two instances of

g

annoy-ing

in

g

The annoy-ing

[Adj]

g

boy is annoy-ing

[

V

]

g

everyon

e

,

while a single

function could have several phonological realizations, as the agentive

nominalization variously realized as -

e

r

,

-

ee

,

and

-ent

in

t

runner

‘one who runs’,

r

standee

‘one who stands’

,

and

correspondent

‘one who corresponds’. The question

t

hen is: are we dealin

g

with three differ

e

n

t phenomena or is there a common thread

t

hat

u

nit

es

th

e

m all

?

W

ithin the framework of LMBM, the problem pertains only to Morphemes; all

L

e

x

e

m

es

ha

ve

a m

o

r

eo

r l

ess

imm

u

ta

b

l

e

o

ne-to-one or one-to-many relationship

with their meanings which precludes zero or empty morphemes. Lexical asymmetry

i

s only possible in cases of (1) synonymy and (2) polysemy, both of which are

i

nconsistent with the facts of asymmetry. Synonyms are so imprecise that most

semanticians ar

g

ue that perfect s

y

non

y

ms do not exist. So

f

a an

d

couch

ma

y

seem to

h

ave identical meanin

g

s, but close exa

m

i

nation demonstrate that the

y

var

y

dialecticall

y

. Morpholo

g

ical modifications, e.

g

., affixes, share identical meanin

g

s

t

hat are unrelated to isoglosses.

Polysemy is also imperfect, usually a matter of metaphoric variation as

cut

c

a

n

m

ean ‘sever’ or ‘insult’. The meanings of polysemantic lexemes are unpredictable

across words, their meanings referring to the real world. Morphological

m

odifications with multiple meanings select from a single pool of meaning, always

grammatical functions. So

morphological

asymmetry is the proper term to use in

l

r

eferrin

g

to the traditional problem of as

y

mmetr

y

.

Roman Jakobson (1939) noted that in a set of morpholo

g

ical relations like the

a

g

entive nominalizations

read

-

e

r

,

stand

-

ee

,

correspond-en

t

,

and

reco

r

d

-

ist

,

where

s

p

eakers know an affix is common, the occa

s

io

nal

o

mi

ss

i

o

n

o

f a

su

ffix

c

an

be

tak

e

n

as a morphological marker itself, and forms like

(

a

)

coo

k

,

guid

e

,

(fast) stud

y

t

herefore become

p

ermissible. The basic fact is that many words across various

l

anguages have a grammatical function without a phonological realization, strongly

suggesting that (1) variation

o

f function

(

derivation

)

and

(

2

)

realization

(

the addition

o

f affixes, prosodic modifications, etc. to the phonemic description of a lexical item)

are separate processes

.

If derivation, the variation of

g

rammati

c

al functions like Tense

,

Person

,

and

N

umber, is a process separate from phonological realization, we would also expect

t

he obverse

p

henomenon: realization without a

n

y functional variation. In fact, this

LEXEME

-

MORPHEME BASE MORPHOLOGY

191

phenomenon is not only common, but one of the age-old conundrums of

m

orphology: empty morphemes. Several empty morphemes are found in English,

e

.g. the extension -

at

in Greek borrowings whose stem ends on

t

-

m

:

d

r

am

-

at

-

ic

(compare

cub

-

ic

,

met

r-

ic

,

and

bas

-

ic

)

or the extra -

al

added to adjectives with this

l

s

ame suffix, e.

g

.

d

ram-at-ic-al

,

metr-ic-al

,

s

y

ntact-ic-al

.

All th

ese wo

r

ds

m

e

an th

e

s

am

ew

ith

o

r

w

ith

ou

t th

e

final -

al

,

so -

al

has no meaning or function.

l

This is crucial evidence supporting the

S

eparation Hypothesis

,

that the rules

which change or adjust the functions in derived words, like

bak-er

from

r

bake

,

op

erate inde

p

endent of the rules that assi

g

n the affix marking it. If the derivation of

baker

from

r

bake

involved only one operation, adding a meaningful suffix

-e

r

,

agentive derivations such as

cook

would have only the meaning of the underlying

k

v

erb and any sentence in which it occurs would be ungrammatical since no suffix

which bears a meanin

g

appears on it.

Morpholo

g

ical as

y

mmetr

y

follows from the Separation H

y

pothesis, too. If deri-

v

ation and phonolo

g

ical realization were ind

e

pendent processes, we would expect

s

ome functional variation to be realized by one realization rule (annoyin

g

,

the verb

,

noun, and adjective) and one function, o

r

functional change, to be realized by

s

everal phonological rules – exactly what we find:

reade

r

,

corres

p

onden

t

,

standee

,

typis

t

. The phenomena of morphology contain exactly what is predicted by LMBM

–

no more

,

no less.

2

.2 The Unitary Grammatical Function Hypothesis

The nature of the categories of derivational morphology (classic ‘word

formation’) was long ignored as morphologists focused on inflection. Beliü

(

1958

)

and Kury

á

o

wicz (1964) first noticed striking parallels between the categories of

i

nflection and derivation in Slavic languages. Bybee (1985), however, re

j

ected the

parallel between the two sets of categories, arguing instead for a continuum between

t

he two sets, with no clear line distinguishing them. In his classic monograph on the

s

ub

j

ect of derivational cate

g

ories, Sz

y

manek (1988) ar

g

ues that derivational

c

ate

g

ories are, in fact, defi

n

able but claims that there are too man

y

of them with

m

eanin

g

s too diverse to be associated with inflection.

B

eard (1995), however, goes through the inflectional categories one by one and

a

ssoc

iat

es

m

os

t

o

f th

e

m

w

ith a

de

ri

v

ati

o

nal category. The categories themselves are

l

isted in the A

pp

endix. Th

e

Unitary Grammatical Function Hypothesis

(

UGF

)

c

laims that the functions that derivational rules operate on are the same for lexical

derivation as for inflectional derivation, e.g. Subject, Object, Possession and

L

ocation. Consider the following example:

(

1

)

The baker bakes cookies.

B

aker

is the subject of this sentence but

r

i

t is also the sub

j

ect of the verb

c

ontained within itself. That is

,

bake

r

,

as a word

,

refers to ‘someone who bakes’ in

t

he sense of the sub

j

ect of the

bake

to which the Sub

j

ective (A

g

entive) -

er

is

r

attached. Notice that one mi

g

ht be tempted to sa

y

that -

er

is the subject of

r

bake

in

1

92

R

O

BERT BEARD

&

MARK

VO

LP

E

the verb, however this explanation of the relationship of the meaning of

baker

to

r

bake

is limited to agentive derivations with phonologically overt affixes. It cannot

apply to all words with analogous relations, e.g.,

cook

,

guid

e

,

and

d

ro

p

-ou

t

,

since no

phonological piece in these nominali

z

ations could bear the sub

j

ec

t r

o

l

e.

It i

s bes

t t

o

s

imply say that the concept of Sub

j

ect i

s

incorporated into the derivation by a

derivation rule which is sometimes marked by a phonological modification of the

l

exical phonolo

gy

, sometimes not. We do not have to offer an alternative source o

f

the Sub

j

ect function in

baker

and

r

cook

.

2

.3 The Base Rule Hypothesis

2

T

o show how syntactic inflectional and l

e

xi

c

al

de

ri

v

ati

o

nal f

u

n

c

ti

o

n

sc

an

be

th

e

s

ame, we must go back to the first modification of Chomsky’s syntactic theory,

so

m

e

tim

es

r

e

f

e

rr

ed

t

o

a

s

th

e

E

xtended Standard Theor

y

(Chomsky, 1965). This

framework posited two syntactic components: (1) a base (categorical) component

and

(

2

)

a transformational com

p

onent. The Base Com

p

onen

t

provides syntactic

t

c

ate

g

ories such as A

g

ent, Patient, Loca

t

ion, Means, Ori

g

in, and Source. LMBM

assumes that all

g

rammatical cate

g

o

r

i

es (includin

g

Tense, Case, Number,

C

om

p

arison

)

and their functions

(

such as Tense: Present, Future, Past; Case:

Subject, Object, Possession, Location, Means, Origin, etc., Comparison: Positive,

C

om

p

arative, Su

p

erlative

)

are located in the Base Com

p

onent.

3

L

et us assume that the sub

j

ect of the sentence in example (1), above, begins its

r

ise to surface structure as shown in

(

2

)

.

I

n (2) we see that a t

y

pical DP stru

c

ture with the function S

U

BJECT could

e

mer

g

e at the surface as

the one who bakes cookies

if

the

,

one

,

and

who

,

i.e.

,

i

ndependent Morphemes in LMBM,

f

ill out the empty nodes. Notice how

f

r

emarkably close the meanings of

one who bakes cookies

an

d

cookie-baker

are.

r

Assume that the morphological categories in the empty nodes in example (2)

above are not recognized by th

e

Lexicon. In this case they would pass unfilled to the

upper levels of grammar. Since they are morphological categories (assuming

Morpheme as defined above), the mo

r

phological component would recognize and

rr

r

ealize them in a syntactic structure as phonological Morphemes, which include

free-standin

g

pronouns, affixes, and other modifications of the stem

,

as determined

b

y

the morpholo

g

ical s

y

stems of particular lan

g

ua

g

es.

2

See Botha (1980), Halle and Marantz (1993), Szymanek

(1985) for other arguments for the Base Rule

k

Hy

pothesis.

3

LMBM assumes Matthews’ (1972) interpretation of

g

rammatical cate

g

ories. Accordin

g

to Matthews,

grammatical categories like Case, Number, Tense, Aspect, Number, Gender each comprise a set of

functions, e.g. Nominative, Accusa

t

ive, Genitive Cases, Singular and Plural Number

,

Past

,

Present

,

and Future Tenses

,

etc. The functi

o

ns, in turn, comprise sets of features

,

as the Genitive Case –

m

arked b

y

the preposition

of

in English – comprise the Genitive of Possession (

f

the propert

y

o

f

the

cit

y

)

, the Partitive Genitive

(

the house of 7 gables), Sub

j

ect and Ob

j

ect Genitives

(

the arrival of the

b

oys

,

the destruction of the cit

y

), among others. To simplify matters, we will combine grammatical

f

u

n

c

ti

o

n

s

an

d

f

e

at

u

r

es

an

d

r

e

f

e

r t

o

both as “functions.” This ste

p

h

a

s

no effect on the theor

y

and does

n

ot conceal an

y

crucial issue of morpholo

g

ical theor

y

.

LEXEME

-

MORPHEME BASE MORPHOLOGY

193

(

2

)

DP

4

D

et

D'

[

+Definite

]

D

C

P

[

Function F

]

IP

DP

V

P

D

V

D

P

[

SUBJECT

][

+Trans

]

bake

D

N

P

[

OBJECT

]

N

cookie

A

dditionally however, LMBM claims that the Lexicon contains rules of its own.

The Lexicon recognizes the morphological categories in (2) but contains no rules fo

r

i

nflectional morphology. It

does

,

however, contain lexical rules operating over the

same categories and their functions.

The Lexicon is the lo

g

ical place for lexi

c

al rules. Lexical rules will var

y

from

l

an

g

ua

g

e to lan

g

ua

g

e but will be constrain

e

d to rules operatin

g

over the cate

g

orical

functions of universal s

y

ntax, such as Sub

j

ect, Ob

j

ect, Goal, and Location. (See

A

ppendix) The only potential output of the Lexicon is words so lexical rules would

h

ave to reduce (2) to a single word without ignoring any of the functional relations

i

n

(

2

)

.

The only option for the Lexicon would be to incorporate the functions of (2) into

an output comprising a single word,

cookie

-

bake

r. It doesn’t matter how the Sub

j

ect

and Ob

j

ect functions are ordered in the lexical description of the output, for the

semantic component will know how to sort

t

hem out. It is onl

y

important that the

semantic component know that the relationship is Sub

j

ect-Ob

j

ect rather than, sa

y

,

Sub

j

ect-Locative, as in

field-worker

or Manner-Ob

j

ect as in

r

fire-brewed

.

All th

e

i

nformation the semantic com

p

onent needs can be derived from structures like

(

2

)

.

Notice that this approach explains the similarity in meaning of

cookie

-

baker

and

r

the one who bakes cookies

without claiming that one is derived from the other.

R

ather, LMBM claims that both are derived from the same underlying syntactic

s

tr

uc

t

u

r

e.

The Unitar

y

Grammatical Function H

y

pothesis, in conclusion, claims that the

functions of derivational (lexical) morpholo

gy

are identical with those of inflectional

m

orpholo

gy

(morpho-s

y

ntax). The functions are inherent in the base structure of the

g

rammar, which is not necessarily syntactic since both higher syntactic rules and

4

Bare phrase structure (Chomsky, 1995) argues for the elimination of the intermediary node X’ in favo

r

o

f th

e

la

be

l

sX

max

(

XP

)

and X

min

(X

n

0

) on minimalist

g

rounds. Where intermedediar

y

nodes arise in this

c

hapter, we emplo

y

the traditional label X’ (Jackendoff, 1976).

194

RO

BERT BEARD & MARK

VO

LPE

l

exical rules operate over it. Languages like Yupik and Algonquian with riche

r

m

orphological systems will use more of thes

e

f

u

n

c

ti

o

n

s

in th

e

i

r

lexical morphology

t

han languages with impoverished morphologies like English and Chinese, which

will realize them mostly in syntax. How

e

v

er, all languages are constrained to the

s

ame universal set of grammatical category functions.

3. TYPES OF LEXICAL

(

L-

)

DERIVATION

LMBM is currentl

y

the onl

y

morpholo

g

ical theor

y

comprisin

g

distinct

c

ompetence and performance theories. It assumes that lan

g

ua

g

e contains means of

c

reatin

g

new words based on unconscious a

n

d productive rules similar to those of

s

yntax. These rules are restricted to lexical

i

tems in the Lexicon, e.g. if a new verb,

s

ay,

smike

entered the language, a panoply of derivations are immediately and

automatically available:

smike

r

,

smike

r

s

,

smikable

,

unsmikabilit

y

,

e

t

c.

Th

ese

ar

e

forms made available by the unconscious creative L-derivation rules of English.

H

owever, the Lexicon also expands its st

o

ck of Lexemes by the addition of new

s

tems, e.g. AID

S

,

r

e

p

,

and

jihad

.

These words are typically consciously added in a

process unlike other

g

rammatical processes (s

y

ntax, morpholo

gy

, phonolo

gy

), which

are alwa

y

s unconscious. AIDS

,

lase

r

,

and

sonar

are the result of conscious efforts to

r

c

reate phrases whose initials letters phonotacticall

y

conform to En

g

lish.

Smog

is the

g

r

esult of an assum

p

tion on the

p

art of a s

p

eaker that

,

if a reference contains two

c

onstituents, it should be given a name

c

om

p

rised of half the name of one

c

onstituent and half the name of the other. These are all conscious, logical processes

and not the result of the unconscious rules of a grammar. It follows that they are

r

ules having to do with how we perform morphology rather than morphology itself.

A

nother propert

y

distin

g

uishin

g

this t

y

pe of derivational rule is that their output

i

s not a variation of a pre-existin

g

Lexeme, as

bake

r

,

b

aker

y

an

d

baking

ar

e

a

g

v

ariati

o

n

o

f

bake

,

but a new Lexeme, an expansion of the lexical stock itself.

Processes like these take

p

lace outside

g

rammar and are thus treated in the

performance theory of LMBM. In this brief section, we will not discuss the LMBM

performance theory, known as Lexical Stock Ex

p

ansion

,

however, details of it may

be found in Beard (1981, 1995, and especially 1987).

3

.1 Competence: Grammatical L-Derivation

LMBM allows four and onl

y

four t

y

pes of L-derivation: (1) Feature

V

alue

S

witches,

(

2

)

F

unctional

L-D

e

r

ivation

,

(

3

)

T

ransposition, and

(

4

)

E

xpressive L-

De

r

ivation

.

All but (4) are determined by the nature of the derivational and

grammatical function systems themsel

v

es; (4) remains a mystery for all

m

orphological theories. Let

’

se

xamin

ee

a

c

h in

de

tail

.

3

.1.1 Feature

V

alue

S

witches

Like most contemporary theories of linguistics, LMBM assumes a Lexicon

c

omprising a catalog of entries made up of l

e

x

ical features.

S

ome features of lexical

LEXEME

-

MORPHEME BASE MORPHOLOGY

195

i

tems are morphological, some are semantic. Evidence across thousands of

l

anguages suggests that these features operate as though they are negatively or

positively marked. For example, the category of Number seems to be two functions,

singular and plural, either of which ma

y be positively or negatively involved in

a

sy

ntactic a

g

reement. For this reason, numbe

r

i

s represented in the lexical entries of

n

ouns as: [±Sin

g

ular], [±Plural]. The noun

table

would be represented as [+Sin

g

ular,

–

Plural] while the Number re

p

resentation fo

r

tables

would be [–Singular, +Plural].

A

curious as

p

ect of this form of re

p

resentation of Number is that it

p

redicts two

o

ther combinations: [+Singular, +Plural] a

n

d [–Singular, –Plural]. What could these

r

epresentations possibly describe?

I

n fact, exactly two more Number phenomena are found in languages which

o

ther theories do not predict and only with difficulty explain: mass nouns (which

h

ave no lo

g

ical or morpholo

g

ical plural, e.

g

. contemplation, env

y,

an

d

sleep

)

and

c

ollective nouns which are both sin

g

ular and plural.

5

It makes no sense to s

p

eak of

nu

m

be

r in

co

nn

ec

ti

o

n

w

ith ma

ss

n

ou

n

s

lik

e

envy

, so the feature value settin

g

[

–

Singular, –Plural] succinctly describes them. Since the default Number is singular,

t

he morphological component will automati

c

ally assign the

m

s

ingular morphology,

w

hat

eve

r that i

s.

I

n Serbo-Croatian, on the other hand, in addition to the plural pera

of

per

o

'

feather

,

' we find a collective for

m

perj

e

,

which agrees in the singular but refers to ‘a

set of feathers’. It is

g

rammaticall

y

sin

g

ular but semanticall

y

plural, hence the

settin

g

[+Sin

g

ular, +Plural]

w

ill

desc

ri

be

thi

sc

la

ss.

6

Man

y

words can have two or three Number forms but one such form alwa

y

s

eems to be the basic or base form. For exam

p

le, it a

pp

ears that

tables

i

sde

ri

ved

f

r

om

table

,

and that the latter is the base form. The base form of

p

era an

d

perja i

s

p

er

o

.

If this assum

p

tion is correct, the Lexicon must contain a rule which converts

s

ingular nouns into plural and collective ones. The only operation required of these

r

ules is the power to change the values of the Number features, as illustrated in (3).

(

3) [+Sin

g

ular, –Plural]

Æ

[–Sin

g

ular, +Plural]

W

e call such rules Feature

V

alue

S

witches to describe the process.

3.

1

.

2 F

unctional

L

exical

-D

e

r

ivation

T

he derivation illustrated by example (2

)

above involves ano

t

her process: the

t

a

dd

iti

o

n

o

f f

u

n

c

ti

o

n

scu

ll

ed

fr

o

m th

e

Ba

se

C

omponent. The addition of values to a

l

exical base requires a discrete type of action by the Lexicon. The single syntactic

base structure of example (2) may cu

l

minate in two possible realizations

m

orpholo

g

icall

y

. The

G

rammatical Functions

p

resent as features in the base

5

Some languages also have a Dual (indicating two ob

j

ects) and a few, a Trial (indicating three). LMBM

takes these to be alternative inter

p

retations of [+Singular, +Plural], since none of these languages have

C

ollective nouns.

6

The Serbo-Croatian collective nouns are similar to

B

ritish collectives which trigger plural number to

v

erbs when the

y

are sub

j

ects, e.

g

. the

g

overnment are in session

,

the team practice ever

y

da

y

,

e

t

c.

196

RO

BERT BEARD & MARK

VO

LPE

s

tructure

,

i.e.

,

SUBJECT and OBJECT

,

assure that their meanings remain

s

ubstantially unchanged.

O

ne result is that the selected Lexemes

bake

an

d

cookie

become

p

art of a

s

yntactic structure whose empty syntactic nodes are completed by the insertion of

Morphemes, resulting in a clause, i.e.,

the

one who bakes

cookies

:

(

4

)

DP

D

et

N

P

[+Definite

]

the

N

C

P

one

spec

IP

[

+Human

]

who

i

DP

i

I

’

D

I

V

P

[SUBJ

]

[Present

]

[Singular]

V

DP

[3

rd

][

+Trans

]

bake

-

s

DN

P

[

OBJ

]

N

Cookies

A

second

p

ossible result is an exam

p

le of a F

unctional

-L

exical

D

e

r

ivation

.

F

unctional Lexical-Derivations consist of a Lexeme combined with one of the forty-

f

our

UG

F

s

r

ecognized by LMBM (see Appendix). The operation of Functional

Lexical-Derivation takes place completely

i

n th

e

L

e

xi

co

n

w

h

e

r

e

it “r

educes

t

o

a

l

exical variety of raising with the amalgamation (incorporation) of complements and

ad

j

uncts” (Beard, 1995: 349).

Th

e

L

e

x

e

m

e

bake

,

incorporated into

t

h

e

h

e

a

do

f an NP

w

ith th

e

f

u

n

c

ti

o

n

SUBJECT

,

becomes the noun

bak

-

er

via the Transposition V

r

ĺ

N

(

see below

)

.

Post-syntactic morphological

sp

ellout of the a

pp

ro

p

riate determiner, the

p

roclitic

of

,

f

t

he

p

lural -

s

and the agentive suffix

-

(

e

)

r

yields

r

the baker of cookies

:

LEXEME

-

MORPHEME BASE MORPHOLOGY

19

7

Further incorporation of the complement

cookie

in the Lexicon, prior to

m

orpholo

g

ical spellout,

y

ields the compound

cookie

-

baker

at spellout:

r

(

6

)

DP

Det

NP

[

+Definite

]

the

N

DP

[

OBJ

][

SUBJ

]

N

P

cookie

k

bake

k

i

-

r

N

t

k

t

T

he most commonly occurring grammati

c

al f

u

n

c

ti

o

n f

o

r n

o

minati

ve

-a

ccus

ati

ve

l

anguages is SUBJECT, a necessary argument

o

f all VPs. For this reason, examples

o

f SUBJECT Functional Lexical-Derivation of the type made possible by (2) abound

i

n En

g

lish, e.

g

.,

boxe

r

,

j

ud

ge

,

and participan

t

.

In accord with the Separation

H

y

pothesis their phonolo

g

ical realizations are varied despite realizin

g

the identical

Grammatical Function

I

n summary, languages choose from the set of universally available Grammatical

Functions

(

the Unitary Grammatical Function Hypothesis) which then may be

r

ealized as inflectional morphology in the

s

yntax or as derivational morphology, i.e.,

Functional Lexical-Derivations in the Lexicon. Volpe (2002) is a specific argument

f

o

r

de

-n

o

minal

LOC

ATI

O

N an

d

LOC

AT

U

M

V

ERB

S

,e

.

g

.,

to shelve

an

d

to saddle

,

as

(

5

)

DP

Det

NP

[

+Definite

]

the

N

DP

[

SUBJECT

]

bake

i

-r

NP

N

C

P

[OBJECT

]

of-cookie

k

-s

k

C

IP

t

i

DP

V

P

D

N

P

V

DP

t

i

t

i

N

D

N

P

t

i

t

k

t

N

t

k

t

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

Подождите немного. Документ загружается.