?tekauer Pavol, Lieber Rochelle. Handbook of Word-Formation
Подождите немного. Документ загружается.
CO
N
S
TRAINT
SO
N PR
O
D
UC
TI
V
ITY
35
1
B
owerman, Melissa. 1982. “Reorganizational processes
i
n lexical and syntactic development.” In: E.
W
anner and L. R. Gleitman
(
eds.
)
, Language Acquisition. The State of the Ar
t
.
Cambridge:
C
ambridge University Press, 319-346.
Dressler, Wolfgang U. 1977. “Phono-morphological dissi
m
i
lation.” In: W.
U
. Dressler and O. Pfeiffe
r
(
eds.
)
, Phonolo
g
ica 1976
.
Inn
sb
r
uc
k
:
In
s
tit
u
t f
ür
Sprachwissenschaft, 41-48.
Fabb, Ni
g
el. 1988. “En
g
lish Suffixation is Con
s
t
rained onl
y
b
y
selectional restrictions.”
N
atural
Language & Linguistic Theory
6,
5
2
7-5
39
.
Gie
g
erich, Heinz J. 1999. Lexical Strata in En
g
lish
.
Cambrid
g
e: Cambrid
g
e Universit
y
Press.
Goldsmith
,
John. 1990. Autose
g
mental and Metrical Phonolo
gy
.O
xford: Blackwell.
Ha
y
, Jennifer. 2002. “From speech perception
t
o morpholo
gy
: Affix-
o
rderin
g
revisited.” Lan
g
ua
ge
78
,
5
27-555.
Hay, Jennifer and Baayen, Harald. 2002. “Parsing
a
n
d productivity.” In: G. E. Booij and J van Marle
(
eds.
)
, Yearbook of morphology 2001
.
Dordrecht: Kluwer
,
203-255.
Hay, Jennifer and Plag, Ingo. (in press). “What
c
onstrains possible suffix combinations? On the
i
nteraction of grammatical and processing
r
estrictions in derivational morphology.”
N
atural
Language & Linguistic Theor
y
.
Kiparsk
y
, Paul. 1982. “Lexical morpholo
gy
and phonolo
gy
.” In: Lin
g
uistics in the Mornin
g
Calm
.S
eoul:
Hanskin
,
3-91.
Krott, Andrea; Schreuder, Robert; and Baa
y
en, Harald. 1999. “Complex words in complex words.”
Lin
g
uistics 37
,
905-926.
L
evin, Beth and Ra
pp
a
p
ort, Malka. 1988. “Nonevent
-er
nominals: a probe into argument structure.”
r
Lin
g
uistics 26
,
1067-1083.
Lieber
,
Rochelle. 1981.
O
n the Organization of the Lexicon
.
Ph
.
D
.
MIT
Malicka-Kle
p
arska, Anna. 1985. “
P
arallel derivation and lexicalist morphology: The case of Polish
diminutivization.” In: E. Gussmann
(
ed.
)
, Phono-Morpholog
y
.
Lublin: Redakcja Wydawnictw KUL,
9
5-112.
Malicka-Kleparska, Anna. 1992. “Ag
a
i
nst phonological conditioning of WFRs.” In: J. Fisiak, Jacek and
S. Puppel (eds.),
P
honological Investigations
.
Amsterdam/Philadelphia: John Benjamins, 423-442.
Marle, Jaap van. 1985.
O
n the Paradigmatic Dimension of Morphological Productivit
y
.
D
o
r
d
r
ec
ht
:
F
o
ri
s.
Marle, Jaap van. 1986. “The Domain H
y
pothesis
:
The stud
y
of rival morpholo
g
ical processes.”
Lin
g
uistics
24, 601
-
62
7.
Menn, Lisa and Mac Whinne
y
, Brian. 1984. “The Repeated Morph Constraint: Toward an explanation.”
Language 60
,
519-541.
Pinker
,
Steven. 1989. Learnability and Cognition
.
Cambridge, Mass.: MIT Press.
Pla
g
, In
g
o. 1996. “Selectional restrictions in En
g
l
ish suffixation revisited. A repl
y
to Fabb (1988).”
Linguistics 34, 769-798.
Plag, Ingo. 1998a. “Morphological haplology in a c
o
nstraint-based morpho-phonology.” In: W. Kehrein
and R. Wiese
(
eds.
)
,
Phonology and Morphology of
the Germanic Languages
f
.
Tübingen: Niemeyer,
199
-
21
5
.
Plag, Ingo. 1998b. “The polysemy o
f
-
i
z
e
derivatives:
O
n the role of sem
a
n
tics in word formation”. In:
G
.
B
ooij and J. van Marle (eds.),
Y
earbook of Morphology 199
7
.
Dordrecht: Kluwer
,
219-242.
Plag, Ingo. 1999.
M
orphological Productivity. Structural Constraints in English Derivation
.
B
e
rlin
/
N
ew
Y
ork: Mouton de Gruyter.
Plag, Ingo. 2002 “The role of selectional restric
t
ions, phonotactics and parsi
n
g in constraining suffix
o
rderin
g
in En
g
lish.” In: G. Booi
j
and J. van Marle (eds.),
Y
earbook o
f
Morpholo
gy
2001
.
D
o
r
d
r
ec
ht
:
Kluwer
,
285-314.
Pla
g
, In
g
o. 2003. Word-Formation in En
g
lish. Cambrid
g
e: Cambrid
g
e Universit
y
Press.
Plank
,
Frans. 1981.
M
orphologische (Ir-)Regularitäten.
A
s
p
ekte der Wortstrukturtheori
e
.
Tübingen: Narr.
R
affelsiefen
,
Renate. 1999.
“
Phonological constraints on English word formation.” In: G. Booij and J.
v
an Marle
(
eds.
)
,
Y
earbook of Morphology 1998
.
Dordrecht: Kluwer
,
225-287.
R
ainer
,
Franz. 1988. “Towards a theory of blocking: The case of Italian and German quality nouns.” In:
G
. Booij and J. van Marle (eds.),
Y
earbook of Morphology 1988
.
Dordrecht: Foris
,
155-185.
R
ainer
,
Franz. 1989. I nomi di qualità nell’italiano contemporane
o
. Vienna: Braumüller.
R
ainer, Franz. 1993.
S
panische Wortbildungslehr
e
. Tübingen: Niemeyer.
R
ainer, Franz. 2000. “Produktivitätsbeschränkungen.” G. Booi
j
, Ch. Lehmann, and J. Mugdan (eds.),
Morphology. An International Handbbok
on Inflection and Word-Formation
k
.
V
olume
1
.
B
e
rlin
/
N
ew
Y
ork: Walter de Gruyter, 877-885.
35
2 FRA
N
Z RAI
N
ER
R
ainer
,
Franz. 2002. “Zum Problem der Suffixwahl i
m
Bereich der Relationsadjektive zu spanischen
m
Schriftstellernamen.” In: D. Kremer
(
ed.
)
,
O
nomastik. Band
V
:
O
nomastik und Lexiko
g
raphie,
D
eonomastik
.
Tübin
g
en: Nieme
y
er, 115-127.
R
ainer, Franz. 2003. “Semantic fra
g
mentation in word-formation: The case of Spanish -
a
z
o
.” In: R. Sin
g
h
and S. Starosta
(
eds.
)
,
E
xplorations in Seamless Morpholog
y
.
New Dehli/Thousand
O
aks/London:
SAGE
,
197-211.
R
ainer, Franz. 2004. “Stud
y
in
g
restrictions on patter
n
s of word-formation b
y
means of the Internet.”
I
talian Journal of Linguistics 15, 131-139.
R
yder, Mary E. 1999. “Bankers and blue-chippers: An account of -er formations in Present-day English.”
E
nglish Language and Linguistics 3
,
269-297.
Scalise, Sergio. 1984. Generative Morpholog
y
.
D
o
r
d
r
ec
ht
:
F
o
ri
s.
Siegel, Dorothy. 1977. “The Adjacency Condition and
t
he theory of morphology.” In: M. Stein (ed.),
Proceedings of the Eighth Annual Meeti
n
g of the North East Linguistic Societ
y
. Amherst
,
Mass.
,
189-
197.
S
ie
g
el, Doroth
y
. 1979. Topics in En
g
lish Morpholo
gy
.
New York:
G
arland Press.
S
okolov, Jeffre
y
L. and Snow, Cathe
r
i
ne E. 1994 “The chan
g
in
g
role of
n
e
g
ative evidence in theories of
l
anguage.” In: C. Gallaway and B. Richards (eds.),
I
nput and Interaction in Language Acquistion
.
C
ambridge: Cambridge University Press, 38-55.
Stember
g
er, Joseph. 1981. “Morpholo
g
ical haplolo
gy
”, Lan
g
ua
ge
57
,
791-817.
Strauss
,
Steven L. 1982. “On relat
e
d
ness paradoxes and r
e
l
ated paradoxes”, Linguistic Inquiry 13
,
694-
700.
W
illiams, Edwin. 1981. “On the notions ‘lexi
c
ally related’ and ‘head of a word’.” Linguistic Inquiry 12,
24
5-
274.
Y
ip, Moira. 1998. “Identity avoida
n
c
e in phonology and morphology.” In: S
.
Lapointe, D. Brentari and P.
F
arrell
(
eds.
)
,
M
orphology and Its Relation to Phonology and Synta
x
.
Stanford
,
Cal.: CSLI
Publications
,
216-246
.
353
Štekauer, P. and R. Lieber (eds.), Handbook of Word-Formation
,
353—373.
©
2005 Sprin
g
er. Printed in the Netherlands.
LEXICALIZATION AND INSTITUTIONALIZATION
PETER H
O
HENHA
US
1. INTR
O
D
UC
TI
O
N
B
oth terms in the main title of this chapter usually refer to what can happen to
words during the course of their ‘life’
after
they have been formed. Despite the
r
i
mplication of ‘after’ and ‘life’, this does not mean that they are relevant onl
y
a
s
purely diachronic concepts. Even though many usages of
lexicali
z
ation
do
in fa
c
t
primaril
y
revolve around diachronic phenomena, such as formal
demotivation
or
semantic idiomati
z
ation
,
the more interesting questions about lexicalization and
institutionali
z
ation
relatin
g
to word-formation theories involve important s
y
nchronic
i
ssues. These include: the nature of the lexicon, the extent to which com
p
lex forms
n
eed or need not/must not be listed in it, what facilitates/prevents the listing of new
c
omplex formations, and the balance of idealization vs. sociological and
psychological reality.
2. LEXICALIZATION
Of the two, this is the more widespread term. However, there isn’t complete
c
onsensus about (the scope of) its mea
n
i
n
g
. In the followin
g
I outline the main
r
eadin
g
s of the term
.
1
2.
1 L
exicali
z
ation in a diach
r
onic sense
T
he least controversial and fairl
y
well understood as
p
ect of lexicalization is
c
ertainly the fact that vocabulary items tend to change over time. While, as Welte
(1996: 79) points out, the terminology in early descriptions in the 1
9
t
h
and early 2
0
t
h
c
entury varied even more than it does t
o
day, the phenomena as such are hardly
rece
nt
d
i
scove
ri
es:
Formal changes that complex words can undergo include phonological
phenomena such as vowel reductions as in
[
mæn
]
>
[
m
n
]
, e.g. in policeman
.
A step
1
There are usages of the term that go beyond the focus of this chapter. In NLP (‘Natural Language
Processing’, in Computer Linguistics), for insta
n
c
e, ‘lexicalization’ is used in a basically
o
nomasiological sense, as the filling with lexical mat
e
r
ial
o
f m
o
r
e
a
bs
tra
c
t
c
o
nce
p
tual structures at
i
ntermediate stages of processing. Lipka (2002: 111
)
mentions similar uses of the term within
Generative Semantics. In some diachronic language studies
,
‘lexicalization’ is used as a kind of
antonym to ‘grammaticalization’, although, as Lehmann (2002) has pointed out, this distinction is not
a
a
q
uite so simple, namel
y
in that
g
rammatic
a
l
ization presupposes prior lexicalization of the
i
tem/construction in question.
354
PETER H
O
HENHA
US
up from such minor chan
g
es are cases such as
breakfast
[
t
b
rei
k
f
#Ö
ff
s
t
]
>
[
brekf
f
f
s
t
]
, or chan
g
es
r
esulting from stress shift, e.g.
admi
r
e
[
d
¥
m
ai
]
–
admi
r
able
[
¥
ædm
r
r
r
bl
]
.
A textbook exam
p
le of extreme formal lexicalization is fo’c’sl
e
,f
ro
m
forecastl
e
,
p
ronounced [
f
n
ff
k
s
l
]
, ‘sleepin
g
quarters under the front deck of a ship’, where the
abbreviated spelling reflecting the pronunciation practically makes it an
unanal
y
sable simplex even in the w
r
i
tten form
(
cf. Li
p
ka 1977, 2002: 113
)
.
C
uriousl
y
, thou
g
h, such lexicalized forms sometimes become remotivated in so-
c
alled spelling pronunciations such as
[
w
eist
k
7
t
]
instead of the older pronunciation
of
waistcoat
as [
t
wesk
t
](
cf. Bauer 1983: 52f
)
.
Formal lexicalization phenomena are not limited to phonolo
g
ical ones. Welte
(
1996: 80) mentions e.g. the dropping of the otherwise obligatory plural morpheme
o
f
wo
r
ds suc
h a
s
t
r
ouse
r
s
or
pyjamas when
p
art of a com
p
ound: trouser
p
ocke
t
,
pyjama to
p
,
while in Swedish we find a reverse
p
henomenon: the
p
lural of
en liten
stad
‘a
s
mall t
ow
n’ i
s
små städer
– and the latter also has to be selected in
r
c
ompounds (in the singular!):
en småstad
(cf. German
d
eine
K
leinstadt
)
o
r
småkaka
‘small cake’
,
‘biscuit’. Furthermore
,
in l
e
xi
c
aliz
ed wo
r
ds we
fin
ds
tr
uc
t
u
r
es
that
would be un
g
rammatical as a freel
y
constructed s
y
nta
g
ma: (
to sco
r
ean
)
own-
g
oal
,
but e.
g
.
*
to have an own house
.
(
Cf. also lexicalized
forget-me-not
vs. *
t
forget me
*
n
ot – don’t forget me.
)
2
Many discussions of lexicalization focus on demotivation and/or loss or addition
o
f
se
manti
c
f
e
at
u
r
es
– an
d
alt
e
rnati
ve
t
e
r
m
s have been used by different scholars,
i
ncluding fossilization (Lyons 1977: 547), petrification
(
Leech 1974: 226
)
, and in
particula
r
idiomati
z
ation
.
Kastovsky (1982a) makes the a
dd
iti
o
nal
d
i
s
tin
c
ti
o
n
be
t
wee
n
s
ystemati
c
lexicali
z
ation
,
such as in the re
g
ular addition of ver
yg
eneral features such as [+
PR
O
FE
SS
I
O
NAL
]
in derivations b
y
means of -
er
(
r
lectu
r
e
r
,
r
e
p
orter
,
w
r
ite
r
)
, and non-
sy
stematic, i.e. trul
y
idiomatic semantic lexicalization.
Aronoff (1976: 19, 43) frequently uses the ter
m
semantic drift
h
e
r
e.
A fam
ous
t
e
xam
p
le of his is
t
r
ansmission
not as a regular action nominalization but as the
technical term for a part of a car – namely
the one that transmits the power of the
y
e
ngine to the wheels
.
3
So this semantic specialization
c
an
be desc
ri
bed
a
s
a
dd
iti
o
n
o
f
s
emantic information (regarding ob
j
ect, source and goal of the verb).
B
auer
(
1983: 55ff
)
finds acc
ou
nt
so
f
se
manti
c
l
e
xi
c
alizati
o
n a
s
a
dd
iti
o
nal
s
emantic information problematic, partl
yb
ecause criteria can be mixed, e.
g
. Lipka’s
(
1977
)
distinction between
l
an
g
ua
g
e-internal chan
g
es as opposed to extralin
g
uistic
changes in cultural background. Examples for the former are
mincemeat
or
t
sweetmeat
(
which
a
r
e
suitable for vegetarians) where the second element goes back
to an older broader meaning, ‘food in gen
e
r
al’
,
but which has later narrowed to its
2
Bauer (1983) makes a few further distinctions re
g
a
r
din
g
, for instance, the so-called combinin
g
forms
(German ‘Fugen’) in Danish and German compounds, the isolation of learned roots such as in English
edible
,
or that of ‘syntactic’ patterns such as ‘verb-object’, now unproductive, in compounds like
pickpocke
t
.
3
I.e. the verbal part of the formation
is
motivated; some morphologists would analyse this as a zero-
he
a
ded co
n
s
tr
uc
ti
o
n
L
EXICALIZATION AND INSTIT
U
TIONALIZATION
355
c
urrent meaning. An example for the latter is
watchmaker
– a profession that these
r
days typically no longer includes the making of watches, only repairs.
B
auer (1983: 57f) also sees a problem in specifying what amount of semantic
i
nformation ma
y
be lexicalized. One of his examples is the exocentric compound
redskin
.
This has two established meanin
g
s, althou
g
h not all speakers ma
y
be
familiar with both: a) a person (now a rather politicall
y
incorrect term) and b) a t
y
pe
o
f
p
otato. This in itself shouldn’t be such a
p
roblem as it is far from unusual for one
form to have more than one meaning and thus more than one lexical entry
(polysemy and homonymy are all over the place in the lexicon anyway). However,
B
auer
(
1983: 58
)
also claims that the relevant additional semantic information is
r
ather down to context than to specializing lexicalization, since
redskin
c
an
be
r
einterpreted with ease as in the following context:
4
(
1) Granny Smith was rude to all lesser breeds, but particularly to M
r
Mackintosh. “You redskins”
,
she would sniff
,
“You’re all alike: no
firmn
ess o
f
c
hara
c
t
e
r
.
”
W
hat Bauer fails to acknowledge here, however, is that this instance of word-
play – a case of a “nonce-use” (cf. Hohenhaus 1996: 133ff) – relies on the deviation
from the lexicalized, conventionally fixed sense of
redskin
(presumably rather the
‘person’ reading), otherwise we would not recognize it as such. Consider also the
m
ore recent example
warhead
. Normally this is lexicalized as ‘the explosive front
e
nd of a missile’. In news coverage of the 2003/2004 demonstrations against the Iraq
war
,
however
,
banners could be seen that had a picture of Geor
g
e W. Bush or Ton
y
Blair next to this word, clearl
y
promptin
g
a different, nonce reinterpretation.
Add
iti
o
n
o
r l
oss
5
of semantic information through lexicalization, and
demotivation, is often referred to as idiomatization, to cover the general aspect of
l
ack of compositionality, which such lexicalized words indeed share with idioms
proper such as
to be ove
r
the moon
‘to be very happy’.
I
t is important to note, however, that idiomatization is only one aspect o
f
l
exicalization, which is why the two terms should not be used interchangeably (as is
sometimes the case
)
. Rather ‘le
x
i
calization’ has to be re
g
arded as the cover term fo
r
a ran
g
e of phenomena, semantic and non-semantic. Bauer (1983: 49) also
emphasizes that “opacity is not a necessary
pre-requisite for lexicalization” since
y
“[s]ome lexicalized forms […] may re
m
ain perfectly transparent”, e.g.
wa
r
mth
–
w
hi
c
h m
us
t
be co
n
s
i
de
r
ed
l
e
x
i
c
aliz
ed bec
a
use
“th
esu
ffix -
th
c
ann
o
t
be
a
dded
synchronically to an adjective to provide a noun”.
F
inally, the various subtypes of diachronic lexicalization phenomena are not
neatly separated issues but frequently overlap, e.g.
h
olida
y
is not only
phonologically different fro
m
h
oly day but also semantically specialized: of the
second constituent onl
y
a feature like [
PERI
O
D
O
F TIME
]
remains, while an
“inferential feature […] of the first constituent, namel
y[
NO WO
RK
]
has become an
4
The exact source is not
g
iven, but it is linked to an advertisement for an apple orchard.
5
An oft-
q
uoted extreme exam
p
l
eo
f l
oss o
f f
e
at
u
r
es
i
s
understand
, where both constituents have
practicall
y
lost all the semantic components of
under
and
r
stand
as free forms.
d
356
PETER H
O
HENHA
US
o
bligatory feature of the whole lexeme” (Lipka 2002: 114). A similar example fro
m
de
ri
v
ati
o
n i
s
i
n
f
amous
,
phonolo
g
icall
y
lexicalized as [
¥
inf
f
f
m
s
]
and semanticall
y
l
exicalized in that it doesn’t mean ‘not famous’ but ‘famous in a negative sense’.
L
e
xi
c
alizati
o
n in th
ed
ia
c
hr
o
ni
cse
n
s
e is clearl
y
a
g
radual affair, both
d
iachronicall
y
(proceedin
g
in successive sta
g
es over time) and in the s
y
nchronic
d
escri
p
tion of the results of such
p
roc
e
s
ses
(
cf. Li
p
ka 2002: 113
)
. Thus we have a
c
line ranging from complete formal and semantic opacity, with the results becoming
similar in status to unanalysable simplex words (e.g.
gospel
), via partial
l
i
diomatization/demotivation, and minor vowel reductions, to even fully transparent
f
o
rm
ssuc
h a
s
wa
r
mth
(
despite the synchronically non-productive suffix).
B
ut why do words undergo diachronic lexi
c
alization in the first place? One
o
bvious and immediatel
y
intuiti
v
el
y
plausible explanation i
s
frequenc
y
of use. Lipka
(1981, 2002: 111) has repeatedl
y
stated this as the main reason. However, Lipka
(
1977: 155,161
)
also
p
oints out that the
p
henomenon of h
y
postatization
6
c
an
be
a
r
eason for lexicalization initially.
7
Hypostatization is a side-effect of the naming-
function of word-formation, whereby the e
x
i
stence of a word seems to imply for
speakers the existence in the real world of a single corresponding ‘thing’ or clearly
d
elimited concept
.
8
Indeed, this could help trigger semantic specialization: making
o
ne out of several readings of a potentially ambiguous formation become fixed, the
w
ord thus becoming a sign in its own right, losing its character of a syntagma
–
w
hich is the
g
eneral default characteristic of lexicalization in the diachronic sense
(Lipka 1977: 156).
2
.2 Lexicalization in a synchronic sense: listing/listedness
The term is itself potentially ambiguous, as
lexicali
z
e
+ -
ation
,
can be inter
p
reted
as a) an action nominalization or b) a result nominalization. In a synchronic sense,
t
hen, the term would correspond to
the process of listing
(the entering of a word in
g
t
he lexicon) in a), while b) corresponds to (the state of)
listedness
, i.e. the property
o
f a word of havin
g
a lexical entr
y
in the lan
g
ua
g
e.
The s
y
nchronic relevance of lexicalization in sense b) has frequentl
y
been
pointed out on various
g
rounds, e.
g
. b
y
Mark Aronoff in connection with the
phenomenon of blocking (e.g. of
stealer
by
r
thief
). He uses it as an argument to
f
f
r
efute Di Sciullo & Williams’
(
1987
)
clai
m
that the lexicon was only about the
6
This term is ori
g
inall
y
taken from Ernst
L
eisi, but Lipka uses it in a wider sense.
7
Another cautionary point can be made here: the general assumption that lexemes only acquire
‘idios
y
ncrasies’ throu
g
h
d
iachronic dri
ft
,
is not necessaril
y
quite such a strai
g
htforward truism.
Herbermann (1981: 334) contests this assumption, claimin
g
that some idios
y
ncrasies are t
y
picall
y
t
here from the start as part of the coinin
g
process, i
.
e. before an item is listed
,
so that if it is listed it
w
ill have to be listed complete with these features alread
y
bein
g
present.
8
In fictional contexts, the hypostatization effect can even be exploited to create ‘reality’ in a fictional
s
ettin
g
, sa
y
, in a science fiction novel, where va
r
ious non-existent ob
j
ects are named (cf. Hohenhaus
1996: 319ff). Here, of course, the h
y
postatiza
t
ion effect does not normall
y
tri
gg
er prope
r
le
xi
c
alizati
on
outside
the fictional context (althou
g
h a few such words have made it into common
v
ocabular
y
throu
g
h the fame of their sources, despite th
e
n
o
n-
e
xi
s
t
e
n
ce o
f th
e
ir
de
n
o
tata in th
e
r
e
al
world, e.
g
.
time
-
machine
,
warp speed
,
b
eam me u
p
– cf. sections 4.4 and 4.5
)
.
LEXICALIZATION AND INSTIT
U
TIONALIZATION
35
7
‘lawless’ (and thus ‘boring’) and their strong hypothesis that listedness is thus totally
i
rrelevant to word-formation theory. In this context, Aronoff (1988: 767) stresses the
asymmetry of the blocking relation “in that the blocking item is in the lexicon” (and
the blocked item isn’t and couldn’t be
)
,
i
.e. without listedness the phenomenon of
blockin
g
cannot be captured, irrespective of whether the listeme is in an
y
wa
y
idiosyncratic or not. (Cf. also the exam
ple of lexicalized but fully analysable
m
m
wa
r
mth
above.
)
Lexicalization in the sense of listing/listedness overlaps with
‘institutionalization’ and will be discussed further in conjunction with that concept
i
n
sec
ti
o
n 3
.
2
.3 The lexicon and theories o
f
word-
f
ormation
W
hile it is a fairl
y
uncontroversial commonl
y
held belief that unpredictable
i
dios
y
ncrasies resultin
g
from lexicalization in the diachronic sense cannot and
s
hould not be generat
e
d by general rules, but hav
e
to be captured piecemeal by
l
exical entries, the wider role of lexicalization in the synchronic sense and the nature
o
f the lexicon are issues of acrimonious debate. Proposals in word-formation theory
generally fall somewhere between two extreme poles; a): a maximally rich lexicon,
i
n th
ese
n
se o
f
:
all
established/existing/etc. words are listed in the lexicon – versus
l
b): a minimal lexicon listing as little as possible. The former has been dubbed the
f
ull-entr
y
model
,
the latter the
impoverished-entry model
b
y
Jackendoff (1975).
l
The assumption that onl
y
as little as possibl
e
should ‘be in’ the lexicon, namel
y
onl
y
what is so idiosyncratic that it cannot be captured by any rules expressing general
r
egularities, is a view that is especiall
y
common
p
lace within ‘lexicalist’ and ‘word-
s
yntactic’ generative frameworks such as Lieber’s (1980), (1992), Selkirk’s (1982)
o
r Di Sciullo & Williams’
(
1987
)
– see Sc
a
l
ise and
G
uevara in this volume. This
assumption is usually argued for on the grounds of ‘conceptual simplicity’ of the
o
verall theoretical model
,
but is also a theoretical necessity for any morpheme-based
m
odel (cf. Aronoff 1988: 768). Occasionall
y
an ar
g
ument in favour of the
i
mpoverished-entr
y
model is put forward which is of a ps
y
cholin
g
uistic nature,
n
amel
y
that listin
g
non-idios
y
ncratic formations alon
g
side idios
y
ncratic items
“
would only encumber the speaker’s memory” (Selkirk 1982: 127). While the
former is indeed largely a theory-internal assumption, ultimately down to “choices
[made] in terms of the
p
riorities
o
f the linguist” (Bauer 1983: 200), the
psycholinguistic argument has to be
j
udged against actual psychological evidence
(
see below
).
However, the central argument for an impoverished-entry model has also always
been contested, e.
g
. b
y
Jackendoff (1997: 124), who points out that its internal lo
g
ic
i
s less convincin
g
than it ma
y
seem: “A fairl
y
standard assumption amon
g
st lin
g
uists
i
s that the impoverished en
t
r
y
theor
y
is ultimatel
y
c
orrect.
[
…
]
However,
[
…
]
although ‘conceptual necessity’ requires that th
e
l
e
xi
co
n
e
n
code w
hat i
s
n
o
t
p
redictable, it does not re
q
uire that the lexicon encode onl
y
what is not
p
redictable.”
W
ord-based theories (in the wake of Aronoff 1976) should be expected to lean
m
ore towards a full-entry model. Aronoff himself remained somewhat undecided on
358
PETER H
O
HENHA
US
t
his theoretical issue
(
cf. Hohenhaus 1996: 159f
)
; however, he offers a
p
ractical
argument, namely that a dictionary “is the
c
l
oses
t
we c
an
co
m
e
t
o
th
e
l
e
xi
co
n
o
f a
n
ative speaker’s language” (Aronoff 1976: 116). Herbermann (1981: e.g. 179) also
argues that dictionaries have always listed more than
j
ust the idiosyncratic
m
inimum. Still, ‘large’ dictionaries (suc
h
as the OED) certainly list a good deal
m
ore than can be expected to
r
eflect an average individual’s
mental lexicon
.
9
O
n the
o
ther hand, as Bauer (2001: 35f) points
o
ut, dictionaries a) represent the
norm
,
i.e.
t
h
ees
ta
b
li
s
h
ed wo
r
ds o
f a
s
peech communit
y
, rather than an individual’s mental
l
exicon, and b) an
y
reference work is nec
e
s
saril
y
outdated, la
gg
in
g
behind the rate of
n
ew formations. So reference to lexicographic lists of lexical entries can at best
s
erve as a working hypothesis (cf. also Hohenhaus 1996: 362). It does, however,
s
eem to cast some doubt on extremely minimalist (morpheme-based) impoverished
entry models.
Such models become particularly doubtful, as Jackendoff (1997: 124) points out,
i
f “we are talking about the mental lexicon, not some mathematically optimal
abstract ob
j
ect, so ps
y
cholin
g
uistic evide
n
c
e is absolutel
y
relevant.” Indeed, the
v
iew that a full-entr
y
model is to be preferred is t
y
picall
y
based on the adequac
y
c
ondition of ps
y
cholo
g
ical
r
ealit
y
(cf. also Me
y
s 1985, Hohenhaus 1998)
.
I
ntegrating a refined version of the full-entry model, Jackendoff (1997) develops
a highly intriguing, heavily psycholinguistically informed theory of tri
p
artite lexical
licensing
(instead of late ‘lexical insertion’). It aims to model parallel processing of
g
phonological, syntactic and conceptual structure, linked by correspondence rules (fo
r
i
nterfacing). The ‘lexicon’ is thus conceived of not as a simple list, but as “a
c
ollection of stored associations amon
g
fra
g
ments of disparate representations”
(
Jackendoff 1997: 108
)
, which is in line with what is known about other mental
processes/representations, onl
y
that “the lexicon ma
y
be unique in its size and utte
r
arbitrariness” (compared e.g. to visual r
e
presentations). In my view, Jackendoff’s is
o
ne of the currently most promising-looking approaches unifying lexicalist
generative theory and a broader psychological picture. We will return to
psycholinguistic issues, as relevant to our topic, in section 4.3.
Štekauer (2000: 3) takes an onomasiological approach rather than a
psycholinguistic one, but by separating a word-formation component from a lexical
c
omponent (and makin
g
the
m
c
ontin
g
ent on the speech communit
y
and its namin
g
n
eeds as the startin
g
point) he also mana
g
es to allow full-entr
y
listin
g
, while still
assumin
g
that word-formation as such is 100% re
g
ular and productive (see Štekauer,
t
his volume
)
.
A
final point that can be made here is that without assuming a full-entry model it
w
ould be difficult for a theory to capture certain crucial conceptional distinctions,
n
otably those between existing
wo
r
ds
vs.
possible words
vs.
n
once-formations –
see
sec
ti
o
n
s
3
.
2 an
d
4
.
1
.
9
While Pinker
(
1994: 150f
)
in this context rather
f
lippantly suggests that the number of entries in a
ff
c
ommercial dictionar
y
ma
y
be deliberatel
y
inflated due to publishers' marketin
g
considerations, he
s
till em
p
hasizes: “[t]he brain seems to be reserving an especially capacious storage space […] for the
m
ental dictionary.”
L
EXICALIZATION AND INSTIT
U
TIONALIZATION
359
3. IN
S
TIT
U
TI
O
NALIZATI
O
N
W
hile ‘lexicalization’ (in one sense or the other) forms part of the terminolo
gy
of
m
ost scholars concerned with word-formation
,
the term ‘institutionalization’ is still
l
ess commonl
y
emplo
y
ed, althou
g
h it has
g
ained a certain currenc
y
over the last 25-
30 years. Broadly speaking, it refers to the stage in the life of a wo
r
d at
(
or from
)
the
t
ransitional point between the status of ex-nonce-formation-turned-neologism (cf.
section 4.1) and that of generally available vocabulary item, i.e. a formation that is
l
isted but not (necessarily) lexicalized in the diachronic sense yet.
3.1 Terminolo
gy
O
ver two decades a
g
o Bauer (1983: 45)
l
amented “ver
y
confused terminolo
gy
”
i
n this area. In an attempt to alleviate this confusion, Bauer (1983: 50) su
gg
ested we
u
se ‘established’ as a hyperonym for lexicalization in the diachronic sense, to which
h
e restricts the use of that term, and institutionalization, thus keeping the latter two
t
erms strictly distinct. Laudable as this clarification may have been, it has not
become universally shared (we have to bear in mind that we are dealing with
n
o
tati
o
nal t
e
rm
s
h
e
r
e
1
0
)
. Neither is the restriction of ‘lexicalization’ to its diachronic
sense generally accepted (cf. section 2.2), no
r
i
s that of ‘institutionalization’ to fully
t
ransparent lexemes onl
y
. Often ‘institutionalized’ and ‘established’ are used
s
y
non
y
mousl
y
.
Li
p
ka
(
2002: 112
)
adds a diff
e
re
nt fa
ce
t
o
f
d
i
s
tin
c
ti
o
n when he sa
y
s that he
ado
p
ts the term ‘institutionalization’ to “stress the sociolinguistic aspects of words”
(see next section). By implication, then, a
s
ynchronic sense of ‘lexicalization’ can be
r
etained, but with a focus on lexicological,
t
heoretical as
p
ects – while ‘established’
m
ay still serve as a cover term for both.
B
oth Bauer (1983) and Lipka (2002) also employ another (competing) set of
t
erms, introduced b
y
Me
y
s (1975, 1985), n
a
m
el
y
the dichotom
y
‘t
y
pe-familiar vs.
i
tem-familiar’. As part of a plea for a fu
l
l-entr
y
model, this places the emphasis on
t
he individual ps
y
cholin
g
uistic aspects involved: a formation is
item-familiar
t
o
a
r
speaker if it is recognized as a particular lexeme, i.e. as one which is not new but
already part of the speaker’s lexicon, whereas new (regular) formations are only
r
ecognized as
type-familiar
but not as listemes in his/her individual mental lexicon
r
(and are therefore analysed, i.e. morphologically processed).
The terminological confusion does not end at the competing definitions already
m
entioned. Various further terms remain in use, e.g.: ‘existing’, ‘actual’ (vs.
‘possible’), ‘occurrin
g
’, ‘recei
v
ed’, ‘in use’, ‘coina
g
e’,
e
t
c.
– all m
o
r
eo
r l
ess
v
a
g
uel
y
equivalent to ‘item-familiar’, ‘institutionalized’ or ‘established’. Whateve
r
notational terms one finall
y
settles for, it is important to relate these to several
p
ers
p
ectives
.
1
0
Cf. Lipka 2002, who makes this point frequently, also with regard to various other terms involved, i.e.
notational terms do not have a sin
g
le ‘correct’ definiti
o
n
,
but “can be defined
d
ifferentl
y
in different
frameworks” (Lipka 2002: 13).
360
PETER H
O
HENHA
US
3
.2 Ideal and real speake
r
s and the speech community
B
auer (2001: 34) emphasizes that a category of existing words i
sc
r
uc
ial f
o
r a
n
otion of productivit
y
“concerned with the potentialit
y
of new formations” – in
o
rder to know what would be new
,
we hav
e
t
o
kn
ow w
hat i
s
n
o
t n
ew.
B
u
t h
e
al
so
o
bserves that this entails severe theoretical
p
roblems, in
p
articular the
q
uestion:
“
existing for whom or what?”
One could argue that an
i
deal s
p
eaker’
s
l
e
xi
co
n
co
ntain
s
all
words of the
l
l
anguage in the broadest sense, i.e. also including all potential words since,
according to Bauer (2001: 35), it is “unclear
how such a construct [as ‘the ideal
r
s
peaker’] could be unfamiliar with a possible word”. This is why he rejects the
n
otion of “existing word” as ‘existing for the
i
deal speaker’. I find that rejection a bit
t
oo sweepin
g
. How the ideal speaker’s mental l
e
xi
co
n i
sco
n
ce
i
ved o
f rath
e
r
see
m
s
t
o depend on the theor
y
makin
g
that idealization. Thus both a minimal
i
mpoverished-entr
y
model and a full-entr
ym
o
del are ultimatel
y
co
mp
atible with the
n
otion of an ideal s
p
eaker. For the forme
r
,
the ideal speaker does not list anything
t
hat is predictable by rule, for the latter the ideal speaker memorizes all words,
i
ncluding compositional ones, that represent the permanent vocabulary of a language
at any given (i.e. equally idealized) point in time. What is more important is to ask:
‘existing/possible
as
what?’ – are we talking about potential vs. actual listemes, i.e.
t
he lexicon, or about the output of wo
r
d-formation processes, independent of
whether or not these later enter the lexicon as well? These two
p
ers
p
ectives should
be ke
p
t se
p
arate – see section 4.2.
B
ut what about real s
p
eakers in real
sp
eech communities? The vocabularies of
i
ndividuals
,
their lexical idiolects
,
will
v
ary quite considerably (beyond a certain
c
ommon core), which is one of the reasons why Bauer (2001: 34), like “most o
f
l
inguistic theory” (but unlike most psychological studies), re
j
ects “the knowledge of
t
he individual as an irrelevancy.”
Lipka (2002: 112), drawing on earlier work by Eugenio Coseriu (1967),
1
1
su
gg
ests that we adopt the notion of
norm
a
s
an int
e
rm
ed
iat
e
l
eve
l
be
t
wee
n
l
an
g
u
e
an
d
parole
–
the former of these Saussurian concepts bein
g
understood as the
‘s
y
stem’ of lan
g
ua
g
e, the latter as its indi
v
idual concrete realizations. ‘Norm’
,
on the
o
ther hand, is understood as the “collective realization of the language system”
(
Li
p
ka 2002: 112
)
, and it is this that accounts for
,
for instance
,
established choices
between systematic alternatives such as
to nationali
z
e
rath
e
r than
*
to nationalify/to
national
,
or ‘habitual disambiguation’ of e.g.
s
leeping pill an
d
h
eadache pill
(
‘for
’
v
s. ‘against’)
.
W
hat is obviously crucial for successful
c
ommunication is shared vocabulary in
speech communities – the latter deliberately put
in the plural here, because what is
t
and what isn’t shared vocabular
y
will
v
ar
y
accordin
g
to di
ff
erent
g
roups o
f
i
nterlocutors which can be collectivel
y
t
aken to constitute speech communities (I
assume ‘community’ to imply a good deg
r
e
e of stability, not just any grouping of
s
p
eakers
)
. The smaller the speech community, the mo
re
it
c
an
d
iff
e
r fr
o
m th
e
11
Cf. also Kastovsk
y
(1982a: 33, 205).