Cognitive Brain Research 6 Ž1997. 37–52
Research report
How the brain processes complex words: an event-related potential study
of German verb inflections
Martina Penke a , Helga Weyerts b, Matthias Gross c , Elke Zander c , Thomas F. Munte
¨
Harald Clahsen e,)
c,d
,
a
Department of Linguistics, UniÕersity of Dusseldorf,
40225 Dusseldorf,
Germany
¨
¨
Department of Clinical Neurophysiology, UniÕersity of Magdeburg, 39120 Magdeburg, Germany
c
Department of Neurology, Medical School of HannoÕer, 30625 HannoÕer, Germany
Department of CognitiÕe Science, UniÕersity of California, San Diego, La Jolla, CA 92093-0515, USA
e
Department of Language and Linguistics, UniÕersity of Essex, Colchester C04 3SQ, UK
b
d
Accepted 4 March 1997
Abstract
Event-related brain potentials ŽERPs. were recorded as German-speaking subjects read verbs in correct and incorrect participle forms.
The critical words were presented in three different versions to three different groups of subjects, as part of a simple sentence, in a word
list, and embedded in a story; for each version separate ERPs were recorded. Three types of verbs were investigated, regulars, irregulars
and nonce verbs. We compared correct regular and irregular participles with incorrect ones; the latter had -(e)n on verbs that actually take
-t participles Ž ) getanz-en., or -(e)t on verbs that require -(e)n Ž ) gelad-et .. For the nonce verbs, we compared participles with the
unexpected -(e)n ending with the expected -t participle forms. The ERP responses were very consistent across the three versions of the
experiment: Ži. incorrect irregular participles Ž ) gelad-et . elicited a left frontotemporal negativity; Žii. incorrect regulars Ž ) getanz-en.
produced no differences to the correct ones; Žiii. nonce verbs were associated with an N400 component but did not show a difference
between expected and unexpected endings. We will interpret these findings with respect to psycholinguistic models of morphological
processing and argue that the brain processes regularly inflected words differently from irregularly inflected ones, the latter by accessing
full-form entries stored in memory and the former by a computational process that decomposes complex words into stems and affixes.
q 1997 Elsevier Science B.V.
Keywords: Event-related potential; Language processing; Left anterior negativity; Regular and irregular inflection; Psycholinguistics; Morphology of
language
1. Introduction
One central issue in psycholinguistic research concerns
the question of whether regularities of language involve
internally represented symbolic rules or whether they can
better be explained in terms of assemblies of stored units.
Consider, for example, past tense formation in English.
Under the latter view, both regular Ž fax-ed, walk-ed, etc..
and irregular past tense forms Ž went, came, saw . are said
to be represented and processed in the same way, by the
storage of items and the creation of connections among
them. On the symbol-manipulation view, however, regular
)
Corresponding author. Fax: q44 Ž1206. 87-2085; e-mail: harald@essex.ac.uk
0926-6410r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved.
PII S 0 9 2 6 - 6 4 1 0 Ž 9 7 . 0 0 0 1 2 - 8
past tense forms are explained differently from irregulars,
by a symbolic rule that instantiates a variable Žthe stem of
a verb. with an instance, e.g. fax, and concatenates that
variable with the -ed affix, yielding faxed. More generally,
the controversy revolves around the question to what
extent the recognition and production of inflected words
involve parsing processes, i.e., the segmentation into stems
and affixes, and to what extent it involves accessing
whole-word based representations stored in memory.
To date, the processing of morphologically complex
words has mainly been investigated by taking reaction-time
measures in various kinds of behavioral tasks Že.g. lexical
decision, priming, naming.. The results of these experiments are discussed controversially with respect to three
processing models, cf. Section 2. Another set of evidence
that might bear on the theoretical debate on morphological
38
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
processing comes from modern brain imaging techniques.
These can be roughly divided in those techniques that
provide high spatial resolution but very low temporal
resolution, and those that provide a comparatively low
spatial resolution but excellent temporal resolution. The
former methods include functional MRI with a temporal
resolution in the order of 10 s and stimulation PET using
radioactively labelled water or butanol with a temporal
resolution of about 1 min, while the latter consist mostly of
electrophysiological techniques, in particular the event-related brain potentials ŽERPs.. The first published functional imaging study of processes underlying regular and
irregular morphology was done by Jaeger et al. w18x. In this
study, the PET technique was applied to a production task
in which subjects had to generate past tense forms of
regular, irregular and nonce verbs in English Žsee
w8,45,46,48x for earlier PET studies on non-inflected
words.. Jaeger et al. w18x found differences in the activation patterns between regular and irregular verbs which,
they argue, support the view that regular past tense forms
are represented differently from irregular ones and involve
the processing of symbolic rules. These conclusions are,
however, not completely borne out, as we will show in
Section 5.3.
For the present study, we adapted the technique of
ERPs to investigate German verb inflection processes.
Following a number of unsystematic attempts to apply the
ERP method to language processing, Kutas and Hillyard
w23x in their seminal paper described an ERP component,
the N400, that was elicited by semantic Ža semantically
inappropriate terminal word. but not by physical violations
Ža sentence-final word written in another font.. Subsequent
research Žsummarized, e.g. in w7,17,20,21,40,51,61,62 x. has
demonstrated that a semantic violation is not a necessary
prerequisite for the elicitation of the N400 response and
that the N400 amplitude varies as a function of degree of
semantic association, contextual constraint, position of the
word within a sentence, priming to name, etc. However,
the violation paradigm has proven to be extremely reliable
in evoking N400 responses. It is not surprising, therefore,
that investigators have relied on the presentation of violations in other language domains. Most research in the
morpho-syntactic domain has been done on violations of
case, number and tense in different languages, English
w1,6,24,33,39,42x, Dutch w14–16x, Spanish w22x, German
w11,31,34,35x, and Turkish w32x. Two ERP effects have
been observed to these kinds of morpho-syntactic violations: a positivity with a latency of about 600 ms variably
called P600 w39,41x or SPS Žsyntactic positive shift w16x.
and an earlier negativity with a left anterior frontal distribution called LAN Žleft anterior negativity w11,19,22x..
While the status of the two effects is still hotly debated, it
is safe to say that morpho-syntactic violations reliably
elicit similar brain responses across different languages,
see Section 5.2 for a discussion of LAN effects.
The violation paradigm has turned out to be one of the
most promising avenues for ERP research on language. In
our research, we have therefore adopted this paradigm to
investigate ERP responses to morphological violations.
2. Models of morphological processing
Psycholinguists have suggested three broad approaches
to capture the processing of morphologically complex
words. First, proponents of full-listing models w2,27x do
not attribute any significant role to the morphological
structure of words in the way they are produced or perceived. Recent connectionist networks w13,26,50,53x are
similar in spirit. In these models, all words are claimed to
be stored in the mental lexicon irrespective of whether or
not the linguist’s analysis yields a morphologically structured representation, e.g. a stem q affix combination.
Second, the full parsing model w57–60x claims that only
stems have entries in the mental lexicon and that morphological variants need to be decomposed in processing
before their stems can be accessed. This model assumes
global affix-stripping mechanisms for processing purposes
which do not necessarily correspond to the morphological
structure of a complex word.
Third, dual-route models w49,25,9,56x claim that our
mindrbrain provides us with two qualitatively different
clusters for treating morphologically complex words, each
with its own representational basis and processing mechanism. In the Pinker and Prince model w49x, these two
clusters are linked to the linguistic distinction between
regular and irregular inflection. In the so-called irregular
cluster, inflected words are represented in terms of lexical
entries, and they are processed by accessing full-form
representations stored in memory, whereas in the regular
cluster inflected words are represented through affixation
rules, and in processing they are decomposed into stem q
affix combinations.
To test these models of morphological processing, we
have investigated two inflectional systems of German,
noun plurals and past participles w4,5,44,65,66x. These two
systems, we think, provide more appropriate grounds for
examining the cognitive status of inflectional rules and the
regularrirregular distinction than the English past tense.
Notice, for example, that 95% of the verbs in English are
regular. Hence, regular past tense forms are much more
Žtype.-frequent than irregular ones Žsee w28x.. Moreover,
only regular past tense forms contain a segmentable affix.
This means that potential differences between forms such
as walk-ed versus came could be effects of frequency
differences andror effects of the presence or absence of an
overt affix, rather than effects of morphological regularity
versus irregularity. In German plurals and participles, however, regularity is not confounded with the presence of an
overt affix and with type frequency. Regulars and irregulars have segmentable endings, and unlike in English, the
regular forms do not represent the majority pattern in any
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
of the two systems. Regular -t participles Ž getanz-t ‘‘danced’’. and irregular -n participles Ž gelad-en ‘‘load-ed’’.
have similar type frequencies 1, and the regular -s plural
Ž Auto-s ‘‘car-s’’. is less frequent than any of the irregular
plurals.
The three morphological processing models mentioned
above make different predictions for our ERP experiments.
If morphologically complex words are stored in the mental
lexicon, as argued by the full-listing model, then we would
not expect to find any regularrirregular differences, i.e.,
violations of regular and of irregular inflection should
elicit similar ERP responses. According to the full parsing
model, all morphologically complex words are decomposed into stem q affix if possible. In German, regular as
well as irregular participle and plural forms have segmentable endings. Thus, incorrect participle and plural
forms of both regulars and irregulars might be expected to
produce violation effects similar to those observed for
morpho-syntactic anomalies in previous ERP studies, i.e.,
a P600 or a left anterior negativity ŽLAN.. In contrast to
the other two models, the dual-mechanism approach w49x
would predict different ERP responses for regular and
irregular inflection. Parsing violation effects should only
occur for incorrect regular affixation, i.e., when the regular
participle ending -t and the plural -s are applied incorrectly, and not for misapplications of irregular inflectional
patterns.
In a previous study w66x, we tested these predictions for
German noun plurals. ERPs were recorded as 18 Germanspeaking subjects read sentences that contained German
nouns in correct and incorrect plural forms as critical
words. We found different ERP responses to regular and
irregular plurals: incorrect applications of -s pluralization
to nouns that normally take irregular plural forms, e.g.
)
Bauer-s instead of the correct form Bauer-n ‘‘farmer-s’’,
were associated with a negativity that had an onset latency
of about 200 ms and a maximum at left anterior temporal
sites. In contrast to that, incorrect regulars, e.g. )Auto-n
instead of the correct form Auto-s ‘‘car-s’’, did not produce a LAN, but rather a low amplitude negativity with a
Ž1. a. durchgetanzt
b. ) durchgetanzen
c. aufgeladen
d. ) aufgeladet
e. angeriltet
f. angerilten
‘‘dance-d through’’
‘‘load-ed on’’
‘‘rilt-ed at’’
39
maximum at central sites. This dissociation, we argued,
confirms the regularrirregular distinction of the dualmechanism model in that violations of the -s plural rule
elicited clearly different brain responses from misapplications of irregular plurals.
To determine whether these findings generalize to other
inflectional systems, we will in the present study report
results from three experiments on German participle formation. We will show that, parallel to the plurals, the
distinction between regular and irregular participles is
associated with different brain responses. Most importantly, violations of the regular -t participle rule elicited a
LAN, just like incorrect -s pluralization did. The general
implications of this finding for the LAN as a marker of
rulerparsing violations will be discussed in Section 5.
3. Methods and materials
3.1. Subjects
Healthy young right-handed native speakers of German
were recruited at the Medical School of Hannover, Germany. There were 20 subjects Ž11 women and 9 men, age
range 21–30 years, mean 24.8. in the sentence experiment,
14 subjects Ž4 women and 10 men, age range 22–33 years,
mean 25.6. in the story experiment and 14 subjects in the
word-list experiment Ž4 women and 10 men, age range
22–37 years, mean 27.2.. None of the subjects participated
in more than one experiment. The data of several additional subjects had to be discarded because of a high
artifact rate leading to the rejection of more than 30% of
the trials.
3.2. Stimuli
The critical stimuli differed with respect to verb type
Žregular, irregular, nonce verb. and in the participle endings, -t vs. -n. This resulted in six conditions Žsee Ž1..; 50
items per condition were tested.
Žregular verb, correct -t .
Žregular verb, incorrect -n.
Žirregular verb, correct -n.
Žirregular verb, incorrect -t .
Žnonce verb, expected -t .
Žnonce verb, unexpected -n.
1
Bybee w3x argued that -t participle suffixation applies to the largest number of verb stems and therefore has a higher type frequency than -n participle
forms. Bybee’s frequency counts, however, are only based on a rather small set of verb types, the 1258 so-called basic verbs Ž‘‘Grundverben’’.. Analyses
of larger sets of verbs showed that -t and -n participles have similar type frequencies w4,28,63x.
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
40
In order to minimize differences between participles of
regular and irregular verbs, we did not include any irregular verbs with vowel changes in their participle forms.
Instead, we only included irregular verbs, such as aufladen
– aufgeladen ‘‘to load on – loaded on’’ in which the
participle stem Žs lad- ‘‘load-’’. is identical to the present
tense stem. Hence, the only difference between the irregular and the regular participle forms used in the experiment
were the two endings -t vs. -n. To generate 50 irregular
verbs without vowel changes in the participle, we made
use of prefix-verb formation, a highly productive process
in German. This process combines a prefix such as durch
‘‘through’’, auf ‘‘on’’, an ‘‘at’’, with a base verb, e.g.
tanzen ‘‘to dance’’, and laden ‘‘to load’’ to form verbs
such as those in Ž1.. Note that all the 300 critical stimuli
consist of such prefixrbase verb combinations. The items
used in the experiments were also controlled with respect
to frequency and similarity. All existing participles and the
corresponding base verbs have low token frequencies Žparticiple forms - 30 w30x; participle forms of base verbs
- 500 w30x; base form of the verbs used - 300 w54x.. To
minimize possible association effects, we made sure that
the participles of nonce verbs used in the experiment did
not rhyme with any existing regular or irregular verb. Two
sets of critical stimuli with 300 participles each were
prepared; verbs that had the -t ending in the first set had
the -n ending in the second set, and vice versa.
The 300 critical words were presented in three different
versions, as elements of simple sentences, embedded in a
story and in a word list. For each version separate ERPs
were recorded.
3.2.1. Sentence experiment
The critical words were presented as the terminal words
within simple declarative sentences of the following form:
Ž2. a. Sie haben die ganze Nacht durchgetanzt
Žthey have danced the whole night through..
b. Sie haben ihre Mobel
¨ schon aufgeladen
Žthey have already loaded the furniture..
c. Er hat seine Schroge
¨ nicht angeriltet
Žhe has not angeriltet Žs nonce verb. his Schroge
¨
Žs nonce noun...
Each sentence consisted of six words. To exclude potential semantic associations, the direct object in sentences
with nonce verbs such as Ž2c. was a nonce word too. The
300 sentences were randomized and presented one word at
a time in yellow letters on a blue background. The words
subtended 1.18 of visual angle in height and between 1.28
and 6.28 in width at the viewing distance of 140 cm.
Throughout the whole experiment a yellow fixation dot
was present in the center of the screen.
The words were shown for 300 ms and the stimulus
onset asynchrony between successive words of a sentence
was 700 ms. The interval between two sentences was 3500
ms. After 10 critical sentences a ‘‘test sentence’’ was
shown in red letters, which either was an exact repetition
of one of the 10 sentences shown before or was slightly
modified, by exchanging one word. The task assigned to
the subjects was to judge whether the test sentence was or
was not an exact repetition of one of the previous 10
sentences and to press one of two buttons accordingly. The
performance on this task was virtually perfect and will not
be discussed further. Short breaks of 1–3 min were given
after 40 sentences.
Subjects were tested in a single session while seated in
an easy chair in a dimly lit experimental chamber. The
subject was observed through an infra-red closed circuit
video-system during the experiment. Instructions were
given to minimize movements and eye-blinks during the
experiment. The total experiment took an average of 150
min to complete including the placement of the electrodes.
3.2.2. Story experiment
The same 300 critical words as in the sentence experiment were used. This time the words were embedded in
Table 1
Brief summary of overall ANOVAs
Measure
Electrodes
Type
Early 250–500 ms
temporal
parasagittal
midline
temporal
parasagittal
midline
111, 222
111, 222, 3
11, 222, 33
11
1
1
Late 500–750 ms
Correctness
T=C
T = C = Antpost
T = C = Hem
T=C=A=H
11, 2, 333
11, 3
1, 33
11
111, 333
11
1, 3
11, 33
1
111, 222, 3
1, 2
not applicable
11, 2
11, 222
not applicable
11, 3
not applicable
1
not applicable
Factors: Type s T Žirregular, regular, pseudo.; Correctnesss C Žcorrect vs. incorrect.; Antposts A, electrode site factor Žanterior to posterior direction.;
Hem s H, hemisphere factor; A = B s interaction of factors A and B. Other main effects and interactions omitted.
1, 2, 3: P - 0.05 for experiment 1 Žsentence., 2 Žstory. and 3 Žlist., respectively.
11, 22, 33: P - 0.01.
111, 222, 333: P - 0.005.
1111, 2222, 3333: P - 0.001.
Sentence
Story
List
Sentence
Story
List
Sentence
Story
List
Sentence
Story
List
Sentence
Story
List
Sentence
Story
List
Experiment
C=A=H
F s, P -
1.27,
1.4,
2.2,
3.84,
0.2,
1.6,
0.69,
0.57,
0.00,
0.85,
0.02,
0.35,
24.6,
17.9,
22.9,
8.9,
2.7,
3.9,
–
–
–
–
–
–
–
–
–
–
–
–
0.0001
0.001
0.0004
0.01
–
–
–
–
–
–
–
–
0.20, –
0.34, –
5.5, 0.01
1.85, –
0.18, –
3.0, –
2.0,
1.03,
2.25,
0.20,
0.6,
1.88,
8.1, 0.008
1.0, –
0.69, –
3.78, 0.04
2.2, –
2.4, –
2.91, –
0.42, –
4.87, 0.05
0.66, –
0.32, –
0.05, –
0.47, –
13.8, 0.003
0.41, –
2.3, –
2.2, –
0.7, –
15.1, 0.001
17.1, 0.002
10.1, 0.007
3.78, –
10.9, 0.006
0.7, –
3.1,
1.9,
2.3,
0.58,
1.7,
0.5,
0.07,
1.22,
0.53,
2.4,
1.6,
2.0,
–
–
–
–
–
–
–
–
–
–
–
–
6.9, 0.003
0.02, –
8.6, 0.009
0.94, –
0.25, –
1.9, –
1.85,
0.04,
1.0,
0.68,
0.01,
0.3,
0.32,
0.05,
0.16,
0.02,
0.03,
0.12,
–
–
–
–
–
–
–
–
–
–
–
–
20.0, 0.0003
12.7, 0.004
24.1, 0.0003
3.91, –
0.18, –
2.11, –
Correctness
F s, P -
C=Hem
F s, P -
Correctness
F s, P -
C=Antpost
F s, P -
Parasagittal
Temporal
3.71, 0.02
1.7, –
2.1, –
3.5, 0.04
0.67, –
2.8, –
3.84, 0.02
0.7, –
0.26, –
0.17, –
0.85, –
0.23, –
1.46, –
0.8, –
1.7, –
4.0, 0.015
0.67, –
0.8, –
C=Antpost
F s, P -
3.44,
2.4,
2.3,
0.26,
0.94,
0.04,
0.06,
1.2,
0.01,
0.43,
0.56,
0.13,
4.1,
3.9,
2.3,
2.1,
9.0,
1.3,
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
0.01
–
C=Hem
F s, P -
0.84,
0.76,
1.7,
1.64,
0.68,
1.6,
0.42,
0.3,
0.18,
0.55,
0.13,
1.2,
1.44,
0.4,
1.7,
0.9,
0.29,
1.8,
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
C=A=H
F s, P -
0.35,
1.0,
2.0,
1.19,
2.6,
0.47,
0.18,
0.05,
0.02,
0.35,
0.06,
0.63,
–
–
–
–
–
–
–
–
–
–
–
–
15.7, 0.0008
8.3, 0.015
13.9, 0.0025
1.93, –
0.13, –
0.92, –
Correctness
F s, P -
Midline
–
–
–
–
–
–
–
–
–
–
–
–
0.48, –
4.0, 0.04
2.5, –
1.91, –
2.8, –
5.2, 0.04
1.87,
0.8,
0.7,
3.8,
1.3,
0.43,
0.13,
0.25,
0.48,
0.49,
0.16,
0.98,
C=Antpost
F s, P -
CorrectnesssC; Antposts A: electrode site factor Žanterior to posterior direction.; Hem s H: hemisphere factor; A=B: interaction of factors and B. Other main effects and interactions omitted.
Degrees of freedom, sentence experiment: Correctness, Correctness=Hemisphere: 1, 19; Correctness=Antpost, C=A=H Žtemporal, midline. 2, 38; Žparasagittal. 4, 76.
Degrees of freedom, story and list experiment: Correctness, Correctness=Hemisphere: 1, 14; Correctness=Antpost, C=A=H Žtemporal, midline. 2, 26; Žparasagittal. 4, 52.
a
Late 500–750 ms
C. Pseudo-Õerbs
Early 250–500 ms
Late 500–750 ms
B. Regular Õerbs
Early 250–500 ms
Late 500–750 ms
A. Irregular Õerbs
Early 250–500 ms
Measure
Table 2
Statistical results for the different word types a
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
41
42
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
Fig. 1. ERPs for the irregular participles. Left side: grand average ERPs for correct and incorrect participles. The layout of the electrodes Žsee scheme in
right lower corner. corresponds to the approximate positions on the scalp. On the right side of the figure the left anterior temporal site ŽF7. is depicted on a
larger scale. In all three experiments incorrect irregular participles are associated with a more negative waveform. This negativity shows slight differences
in distribution across experiments but the left anterior temporal site displays the effect in all cases.
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
three stories: an animal fable Ž1312 words total, 104r300
critical words., an episode from the Odyssey Ž1486 words
total, 102r300 critical words. and the tale of the Cologne
43
dwarfs Ž1349 words total, 94r300 critical words.. Unlike
the sentence study the critical words in this experiment did
not always appear clause-finally; they were rather some-
Fig. 2. ERPs for regular participles. Same layout as Fig. 1. For the incorrect participles no consistent difference across experiment is found.
44
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
times followed by a dislocated prepositional phrase, a
grammatically well formed word-order variation in German. Two different versions of the stories were prepared
with correct and incorrect participle forms exchanged. Half
of the subjects were exposed to the first version, the other
half to the second one. The mode of presentation Žcolours
Fig. 3. ERPs for participles of nonce verbs. Same layout as Fig. 1. No consistent differences between participles with the expected and the unexpected
ending emerged across experiments.
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
of words and background, viewing distance. was identical
to the sentence experiment; each word was presented for
300 ms, the stimulus onset asynchrony between words was
800 ms. Subjects were told that they participated in a
memory experiment and after each of the stories they
received a questionnaire with 10 simple questions pertaining to the contents of the story. Performance on the
questionnaires was flawless and will not be discussed
further. The subjects were tested in a single session taking
about 180 min including electrode placement.
3.2.3. Word-list experiment
The critical stimuli were the same as in the sentence
and the story experiments. The 300 critical participle forms,
300 additional verb forms Ž100 infinitives, 100 present
tense forms, 100 past tense forms. used as filler items, and
100 nouns and 20 nonce nouns were assembled in random
order to yield a word list. The word list was presented one
word at a time in yellow letters against a blue background
Žword duration 300 ms, stimulus onset asynchrony 2000
ms.. The subjects’ task was to press a button held in the
right hand whenever they encountered a noun. While in
written German the first letter of nouns has to be upper
case, all stimuli of the list were presented in lower case
letters to prevent subjects from making the nounrverb
distinction simply on the basis of case. Performance under
these conditions was very good Ž) 90% correct responses
to nouns in each case. indicating that the subjects actually
read the stimuli in order to make the required distinction.
Subjects were tested in a single short Žapprox. 70 min.
session.
3.3. Recording and analysis
EEG was recorded from all 19 standard scalp sites 2 of
the 10r20 system using tin electrodes mounted in an
electrode cap with reference electrodes placed at the mastoid processes. Additional electrodes were affixed at the
right external canthus and at the right lower orbital ridge to
monitor eye movements for later off-line rejection. The
biosignals were amplified with a bandpass from 0.01 to 70
Hz, digitized at 250 pointsrs and stored on magnetic disk.
After artifact rejection by an automated procedure using an
individualized amplitude criterion on the eye channel and
the frontopolar channels, ERPs were averaged for 1024
epochs, including a 100 ms prestimulus interval. The
waveforms were quantified by mean-amplitude measures
in time windows 250–500 ms and 500–750 ms. These
data were subjected to repeated measures analyses of
variance. Since effects were differentially distributed over
the scalp, separate analyses were done for the midline
2
Ten additional electrodes were used in the sentence experiment.
However, for the sake of consistency across experiments only the 10r20
sites will be discussed in this paper.
45
Fig. 4. Comparison of participles of nonce verbs with the expected Ž -t .
ending and correct regular and irregular participles for the Cz site. The
nonce verb participles exhibit a negativity that corresponds to the N400
component.
ŽML; Fz, Cz, Pz., parasagittal ŽPS; Fp1r2, F3r4, C3r4,
P3r4, O1r2., and temporal ŽTE; F7r8, T3r4, T5r6.
electrodes with the latter two sets split into an electrode
site Žanterior to posterior, PS: 5 levels, TE: 3 levels. and a
hemisphere factor Žleft vs. right hemisphere.. The analysis
was conducted in two steps: first, an overall ANOVA was
computed with Word Type Žregular vs. irregular vs.
pseudo., Correctness Žcorrect vs. incorrect., Site and
Hemisphere as factors. As significant interaction effects
involving the Word Type and Correctness factors were
obtained in these analyses, additional ANOVAs were conducted separately for each word type to further delineate
the pattern of results. The Greenhouse–Geisser correction
for inhomogeneities of covariance was applied whenever
applicable. Reported P values are corrected.
4. Results
The pattern of results was very similar across the three
experiments; the data from these experiments will therefore be presented together organized according to the
different participle forms, regulars, irregulars and nonce
Fig. 5. Comparison of regular and irregular participles according to
ending for the F7 site: only the incorrect irregulars Žthose with the -t
ending. are associated with the left anterior temporal negativity in all
three experiments.
46
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
forms. Table 1 shows a summary of the results of the
overall ANOVAs. For all experiments and for both time
windows, interaction effects of the factors Type and Correctness were found, as well as interactions of these two
factors with the topography factors Site and Hemisphere.
These interactions warranted the computation of separate
ANOVAs for the different word types Žsee Table 2A–C.
to elucidate their nature; these will be assessed in the
following.
4.1. ERPs to irregular words
The grand-average ERPs to the irregular words are
shown in Fig. 1. The ERPs begin with an ensemble of
early peaks and troughs Žnegativity at about 100 ms,
positivity at about 200 ms. that differed considerably
across the experiments. The differences in these early
exogenous components are due to the physical conditions
in which the stimuli were presented, most notably to the
different inter-stimulus intervals. These were responsible,
for example, for the higher amplitude occipital N1 in the
word-list experiment.
More importantly, the ERPs in Fig. 1 show that in all
three experiments, starting at approximately 150–200 ms,
irregular verbs with incorrect -t endings were associated
with a more negative waveform. This negativity differed
somewhat in its distribution across the experiments but
showed a preponderance at the left frontotemporal ŽF7.
site, which is marked by boxes on the left of Fig. 1. At the
right of Fig. 1, the F7 site is plotted on a larger scale.
Statistically Žsee Table 2A., this negativity gave rise to a
main effect of the factor Correctness for all three electrode
sets across the three experiments for the mean amplitude
measure in the 250–500 ms time window. For the temporal electrode set, interactions of Correctness with the factors Site Žanterior to posterior. and Hemisphere were observed indicating a circumscribed topographical distribution of the negativity. While the effect appeared to extend
into later portions of the waveform, statistical effects for
the 500–750 ms time window were less pronounced.
4.2. ERPs to regular words
The ERPs to the regular verbs ŽFig. 2. showed a similar
general waveshape as the irregular verbs. However, no
consistent differences between correct and incorrect verbs
were seen in the three experiments. In the sentence experiment, the ERPs to incorrect verbs appeared to be slightly
more positive starting at approximately 200 ms, an effect
that was not present in the story and word-list experiments.
Fig. 6. Illustration of the scalp distribution of the negative effect: depicted are the mean differences between correct and incorrect participles in the
250–500 ms time window Žaligned for the prestimulus baseline y100 to 0 ms.. TEMP, temporal electrodes; PARA, parasagittal electrodes. A remarkable
similarity across experiments emerges with the irregulars being associated with a negativity, which has a left sided preponderance, particularly for the
temporal electrodes. The difference between correct and incorrect regular participles words is close to zero in all cases.
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
Statistically Žsee Table 2B., the positive-going waveform
in the sentence study gave rise to a Correctness by Site
Žanterior to posterior. interaction for the parasagittal electrode set in the early time window. In the story experiment, however, a more negative-going ERP was found at
T3 which led to a Correctness by Hemisphere interaction
for the temporal electrode set in the early time window.
Hence, there were no consistent ERP effects for incorrect
regulars across the three experiments.
47
4.3. ERPs to nonce words
The ERPs to the nonce verbs are shown in Fig. 3. With
regard to the two participle forms Žexpected -t versus
unexpected -n., no consistent differences were obtained
across the three experiments. Slightly more positive ERPs
to the -n participles in the word-list experiment at the T5
and T6 sites gave rise to a Correctness by Site Žanterior to
posterior. interaction for the temporal electrode set in the
Fig. 7. A: difference waves Žincorrect minus correct. for the three verb types Žregular, irregular, nonce. averaged across all three experiments Ž n s 48..
Only the difference wave for the irregulars shows a monophasic negativity with a left anterior maximum. B: isovoltage maps using spherical spline
interpolation. Depicted is the mean amplitude in the 250–500 ms time window measured on the incorrect–correct difference waves for the irregulars. The
isovoltage maps show a circumscribed left anterior negativity. Scaling: min, y2.7 mV; max, 1.25 mV. C: current source density maps on the same data as
in B. These maps provide a reference-free estimate of current flow through the skull. They are characterized by a prominent Žnegative. sink at F7 and T3.
Scaling: min, y17.75 mVrm2 ; max, 9.00 mVrm2 .
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
48
early time window. In the sentence experiment, on the
other hand, a slightly more positive waveform for -n
participles was observed at Fp1 and Fp2 causing a Correctness by Site interaction for the parasagittal electrode set.
Recall that the nonce word items used in our experiment are pronounceable words for German native speakers
and that previous ERP studies found an N400 component
especially for pronounceable nonce words w52,37x. To determine whether this also holds for our nonce word items,
ERPs to correct irregular and regular participles were
compared to nonce participle forms with the expected -t
suffix for the Cz electrode Žsee Fig. 4.. For all three
experiments a negativity was observed for the nonce words
that was more prominent in the sentence and story experiments than in the word-list study.
4.4. The distribution of the ERPs to regular and irregular
words
In order to assess the distribution and the extent of the
effects associated with incorrect participles, three analyses
were performed. First, we compared the ERP responses to
regular and irregular participle forms 3 at the F7 site Žsee
Fig. 5.. The waveforms show that only the incorrect
irregular participles are associated with an enhanced negativity, whereas incorrect occurrences of the participle -n on
regular verbs produced ERPs similar to those of the correct
regular participle forms.
Second, we compared the difference in mean amplitude
between correct and incorrect participles within the 250–
500 ms time window for the temporal Župper row. and the
parasagittal Žlower row. electrode sets Žsee Fig. 6.. A
remarkable similarity between the three experiments
emerges: the negativity associated with incorrect irregular
verbs shows a left preponderance in all three experiments
especially for the temporal set. Moreover, the negativity is
most prominent for left frontotemporal and frontal sites.
Note also, that the difference between correct and incorrect
regular participles is close to zero in all cases.
Finally, to obtain a more precise localization on the
scalp, the difference waves Žincorrect–correct. were pooled
from all three experiments. This averaging across experiments was done simply for illustrative purposes with the
idea that the signal-to-noise ratio would be greatly enhanced, thus affording us a better idea of the generalizability of the effects. We felt justified to do this meta-analysis
as the critical items were the same in all three experiments,
the recording parameters were identical and the subjects
were from the same pool. Fig. 7A presents the difference
waves for regular, irregular and nonce verbs, Fig. 7B the
isovoltage maps and Fig. 7C the current source density
maps for the difference between correct and incorrect
3
ERPs to nonce words had to be excluded from Fig. 5, as these were
independently shown to produce an N400 component.
irregular participles. The difference waves show that it is
only the incorrect occurrence of -t on irregulars that elicits
a clear ERP response, namely a monophasic negativity
with a left anterior maximum, whereas the incorrect regulars and the nonce verbs produce a flat line. The isovoltage
maps indicate that the ERP response we found for incorrect irregular participles exhibits a clearly circumscribed
distribution. This is further corroborated by the computation of the current source densities, which provides a
reference-free estimate of the currents passing through the
skull, and suggests a left frontal source for the negativity.
5. Discussion
The most salient and consistent result of the present
study is that incorrect irregular participles were associated
with a negative ERP with a left anterior temporal distribution starting at about 200 ms, an effect that has been called
LAN. The LAN for incorrect irregulars was found in each
of the three participle experiments presented in this study.
Even more strikingly, this brain response to incorrect
irregulars generalizes to an entirely different inflectional
system, noun plurals. We found that incorrect irregular
noun plurals of German elicited a LAN w66x just like
incorrect irregular participles. Note that noun plurals differ
from participle forms, both in terms of structural properties
and in terms of vocabulary distribution. Whereas the regular participle ending -t is added to verb stems, e.g. tanz‘‘danc-’’, kauf- ‘‘buy-’’, leg- ‘‘put’’, the regular plural
ending -s is added to word forms such as Auto ‘‘car’’,
Pizza, Bar, etc., i.e., to elements which exist as independent words in the language. Moreover, the frequency
distribution and the similarity patterns in these two inflectional systems are quite different. Irregular verbs do not
outnumber regular ones, independent of the method of
frequency calculation chosen w64x, but 92% of the German
nouns form their plurals according to one of the various
irregular patterns. Despite these linguistic differences, the
ERP response to incorrect irregulars is very similar for
noun plurals and participles, indicating the stability, reproducibility and generalizability of this negativity.
5.1. ERPs and morphological processing models
The current set of ERP results demonstrate clear electrophysiological differences between regular and irregular
inflection for noun plurals and participles, a finding which
is directly relevant to the competing models of morphological processing introduced in Section 2. Recall that according to the full-listing model regular and irregular inflection
are processed in similar ways, i.e., by the storage of items
and the creation of connections in the mental lexicon.
From such a model, one would predict violations of regular and of irregular inflections to elicit similar ERP responses. Our results show that this prediction is not borne
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
out. The second variant of a single-mechanism approach is
the full parsing model according to which all morphologically complex words are decomposed into stem q affix if
possible. As German participle and plural forms have
segmentable endings, we would again expect to find similar brain responses to incorrect forms of both regulars and
irregulars. This is not confirmed, however. Hence, our
ERP results do not appear to support any of the singlemechanism models.
The dual-mechanism model, on the other hand, would
be consistent with different ERP responses for regular and
irregular inflections. This model distinguishes between two
different clusters for morphologically complex words, the
irregular cluster involving the retrieval of lexical entries
stored in memory and the regular cluster involving morphological parsing. Consider, from the perspective of the
dual-mechanism model, the specific linguistic conditions
tested in our experiments. Two kinds of morphological
anomalies were investigated: Ži. regularizations, -t affixes
appearing on irregular verbs Ž ) geladet .; Žii. irregularizations, the participle -n on regular verbs Ž ) getanzen. or
nonce verbs Ž ) gerilten.. The structure of these morphological anomalies is quite different. The first case is a parsing
violation, i.e., a misapplication of the -t affixation rule to a
verb that would normally block the rule, to produce illegal
stem q affix combinations such as ) gelad-et. The other
two cases are incorrect or unexpected -n participle forms
of verbs that would normally undergo the -t affixation rule.
According to the model, there is no -n suffixation rule in
German participle formation; irregular participles are rather
said to be stored in memory together with the -n ending.
Hence, it is only regularizations that involve rule violations. In such cases, we found a characteristic ERP response in the three participle experiments and the plural
study w66x: the LAN occurred when the regular participle
ending -t and the plural -s were applied incorrectly, but
not in cases of irregularizations. Thus, the LAN found
under these conditions can be interpreted as reflecting
processes involved in Žmorphological. structure building.
5.2. A comparison with left anterior temporal negatiÕities
from other ERP studies
In previous studies, several kinds of morpho-syntactic
anomalies have been claimed to elicit a LAN, e.g. violations of syntactic agreement, phrase-structure violations,
illegal filler-gap constructions and certain kinds of grammatically well formed embedded clauses. Some researchers have interpreted the LAN as reflecting working
memory operations Žw6,19,22x.. From a linguistic perspective, however, the LAN appears to vary as a function of
processes involved in morpho-syntactic structure building.
The LAN occurs when rules of affixation Že.g. subject-verb
agreement. and phrase-structure building are incorrectly
applied and in cases of complex syntactic dependencies
that require extra parsing effort, such as filler-gap constructions and re-analysis of theta role assignment. Our
49
findings on regular and irregular inflections are compatible
with this interpretation; a LAN was found for violations of
regular inflectional rules, i.e., for cases that involve morphological parsing processes, and it was not associated
with violations of irregular inflectional patterns, which do
not involve affixation rules.
Munte
et al. w33x and Adamson et al. w1x found that
¨
violations of subject-verb agreement in English, e.g. ) you
goes and ) your spend, were associated with a more
anterior and left preponderant negativity, while in semantic
judgements pairs of nouns such as sky–thief compared to
robber–thief produced an N400 component. The negativity in the agreement condition was shown to be distinct
from the N400. Most probably, it is related to the fact that
agreement violations involve parsing processes and are
structurally more complex than simple noun pairs. Osterhout and Mobley w38x also found a LAN for violations of
subject-verb agreement but not for incorrectly inflected
reflexive pronouns, such as himself instead of herself or
themselÕes instead of himself. They take subject-verb
agreement to be more syntactic than number or gender
errors on pronouns and argue that the LAN reflects this
difference.
Left anterior negativities were also found in cases of
phrase-structure violations, such as ) Ted’s about films
America w36x, and German sentences such as ) Der Freund
wurde im besucht ‘‘The friend was in visited’’ w10,12x.
The latter authors argue that the LAN is associated with a
relatively low-level early stage of structure building.
Some ERP studies have discovered a LAN in illegal
filler-gap constructions, but results are less consistent
across studies than for agreement and phrase-structure
violations. Kutas and Kluender w22x found a LAN for
ungrammatical wh questions such as ) What can’t you
remember who he adÕised .... This negativity was interpreted as reflecting working memory operations associated
with the filler-gap construction. Such constructions were
claimed to put a strain on the limited capacity of the
working memory system, because the system must store
the filler along with intermediate material. However, results from other ERP studies on the same phenomenon
have produced different results. Neville et al. w36x found a
LAN for so-called subjacency violations ( ) What i was
sketch of — i admired by the man), but not for other
violations of wh movement such as ) What i did the man
admire Don’s sketch of — i , and McKinnon and Osterhout
w29x elicited a P600, but no LAN in such cases.
Summarizing, we suggest a linguistic interpretation of
the results mentioned above in terms of morpho-syntactic
structure building, with the LAN being elicited by violations of morpho-syntactic rules or in cases of complex
syntactic dependencies. The question, however, to what
extent the LAN reflects processes at the level of the
working memory requires further study of linguistic and
non-linguistic stimuli; we will have to leave this question
open.
50
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
5.3. Which brain structures are inÕolÕed in the processing
of inflected words?
If we accept the negativity obtained for incorrect -t
participles and incorrect -s plurals as a correlate of the
processing differences between regular and irregular words,
the question arises as to which brain structures might be
involved in its generation. Recall that the difference waves
and the maps shown in Fig. 7 suggest a left frontal source
of the negativity found for regularizations, i.e., for violations of the -t affixation rule. This might in turn fuel
speculations that it is generated by the same brain areas
which are traditionally said to cause agrammatism. Since
we cannot be sure, however, that the activity measured at
some electrode site is generated by the brain tissue directly
below the electrode, caution must be exerted in advancing
such an interpretation. Clearly, experiments using more
electrode sites, a larger number of stimuli and the employment of the emerging new localization techniques w55,43x
andror magnetoencephalographic recordings are needed to
warrant further reasoning along these lines.
The question as to whether similar or different brain
structures are responsible for irregular and regular inflection has also been the subject of a recent PET study w18x
on the English past tense. While lying in the PET camera,
nine male subjects were performing two different tasks: Ži.
reading aloud of visually presented stems of regular, irregular and nonce verbs; and Žii. producing past tense forms
of these verbs. Using radioactively labelled water it was
possible to assess rCBF differences as a function of the
tasks. For this purpose, subtractions were performed of the
rCBF differences between the past-tense production and
the stem-form reading tasks for the three kinds of verbs.
Jaeger et al. w18x found different areas of cortical activation
in the regular and the irregular tasks: the left dorsolateral
prefrontal cortex ŽBrodmann’s area 46. and the left anterior congulate gyrus Žarea 24. showed activation in regular
and nonce word past-tense production, while the mid-temporal gyrus Žarea 21. and the left superior frontal gyrus
Žarea 10. were active in the irregular past tense condition,
the latter also being active in the production of nonce past
tense forms. Moreover, Broca’s area Ž44, 45. was active in
all three production tasks. Jaeger et al. compared their
activation patterns with other neuropsychological data and,
on the basis of that, they made claims about the localization of past-tense formation processes in specific brain
structures. Broca’s area, for example, is said to process the
tense feature Žqpast., area 46 computes regularly inflected
forms, and area 21 is responsible for the retrieval of
irregular forms. They conclude that the different activation
patterns found for regulars and irregulars provide evidence
against single-mechanism models such as the connectionist
approach and support the dual-mechanism view that regulars are processed differently from irregulars.
Evaluating these far-reaching conclusions should be the
subject of a further study involving a detailed assessment
of the methods and materials used for their PET scans
which we will not address here. One comment, however,
seems to us to be pertinent with respect to the Jaeger et al.
study w18x. Whereas most of their claims are about processing aspects of inflection, particularly the computation of
regulars versus the retrieval of stored irregulars and the
brain structures involved in these processes, the empirical
evidence Jaeger et al. provide from their PET study does
not bear on rapid processes, such as past tense formation,
which take place within 200–400 ms. This is due to the
poor temporal resolution of PETs. What Jaeger et al.’s
rCBF patterns show instead are cortical activations while
subjects produce a list of past tense forms. Leaving this
criticism aside, Jaeger et al.’s data do nonetheless suggest
that regular and irregular inflections are represented in
cortically distinct areas, thus providing support for the
view that regulars and irregulars are based on different
linguistic representations. To address issues of morphological processing, however, ERP experiments such as those
presented here appear to be more revealing than PET
studies.
6. Conclusion
We conclude that two complementary mechanisms coexist in how our brain processes morphologically complex
words: Ži. accessing full-form entries stored in memory;
and Žii. a computational system that decomposes complex
words into stems and affixes. This distinction is reflected
in different ERP responses to Ži. and Žii..
Acknowledgements
Thanks are due to Berdieke Wieringa, Marta Kutas,
Gary Marcus, Sonke
Johannes, Mike Matzke, and Jonathan
¨
King for comments, Jobst Kilian for excellent technical
assistance and Sam Featherston for correcting our English.
T.F.M. was supported by a grant from the Hermann and
Lilly Schilling Foundation ŽTS 013r177r96., and H.C. by
a German Science Foundation grant ŽSFB 282rProject
B7..
References
w1x A.M. Adamson, T.F. Munte,
G.R. Mangun, Differences in timing
¨
and distribution of semantic and syntactic incongruency effects
demonstrated by multichannel ERP recording, in: T.F. Munte,
H.-J.
¨
Heinze, G.R. Mangun, H. Scheich ŽEds.., Mapping Cognition in
Time and Space, IOS Press, Amsterdam, 1997, in press.
w2x B.L. Butterworth, Lexical representation, in: B.L. Butterworth ŽEd..,
Language Production, Vol. 2, Academic Press, London, 1983, pp.
257–294.
w3x J.L. Bybee, Regular morphology and the lexicon, Lang. Cogn. Proc.
10 Ž1995. 425–455.
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
w4x H. Clahsen, The representation of German participles in the German
mental lexicon: evidence for the dual-mechanism model, in: Yearbook of Morphology, Kluwer, Dordrecht, The Netherlands, 1996,
pp. 1–31.
w5x H. Clahsen, S. Eisenbeiss, I. Sonnenstuhl, Morphological structure
and the processing of inflected words, Theor. Ling., 22 Ž1997. in
press.
w6x S. Coulson, J.W. King, S. Coulson, Expect the unexpected: event-related brain response to morphosyntactic violations, Lang. Cogn.
Proc., submitted.
w7x I. Fischler, G.E. Raney, Language by eye: behavioral and psychophysiological approaches to reading, in: M.G.H. Coles ŽEd..,
Handbook of Cognitive Psychophysiology: Central and Autonomic
Nervous System Approaches, Wiley, Chichester, 1991, pp. 511–574.
w8x R.S.J. Frackowiak, Functional mapping of verbal memory and language, Trends Neurosci. 17 Ž1994. 109–115.
w9x U. Frauenfelder, R. Schreuder, Constraining psycholinguistic models
of morphological processing and representation: the role of productivity, in: Yearbook of Morphology, 1991, pp. 165–183.
w10x A.D. Friederici, A. Mecklinger, Syntactic parsing as revealed by
brain responses. First-pass and second-pass parsing processes, J.
Psycholing. Res., 25 Ž1996. 157–176.
w11x A.D. Friederici, A. Hahne, A. Mecklinger, Temporal structure of
syntactic parsing: early and late event-related brain potential effects,
J. Exp. Psychol.: Mem. Cogn. 22 Ž1996. 1219–1248.
w12x A.D. Friederici, E. Pfeifer, A. Hahne, Event-related brain potentials
during natural speech processing: effects of semantic, morphological, and syntactic violations, Cogn. Brain Res. 1 Ž1993. 183–192.
w13x M. Gassar, Learning words in time: towards a modular connectionist
account of the acquisition of receptive morphology, Unpublished
manuscript, Indiana University, Bloomington, IN, 1993.
w14x T.C. Gunter, S.H. Vos, G. Mulder, Syntactic violations and ERPs:
P600 or P3b?, Poster presented at the CUNY Sentence Processing
Conference, Tucson, AZ, 1995.
w15x P. Hagoort, C. Brown, Brain responses to lexical ambiguity resolution and parsing, in: C. Clifton, L. Frazier, K. Rayner ŽEds..,
Perspectives on Sentence Processing, Erlbaum, Hillsdale, NJ, 1994,
pp. 45–80.
w16x P. Hagoort, J. Groothusen, C. Brown, The syntactic positive shift
ŽSPS. as an ERP w15x measure of syntactic processing, Lang. Cogn.
Proc. 9 Ž1993. 439–483.
w17x S.A. Hillyard, T.W. Picton, Electrophysiology of cognition, in: F.
Plum ŽEd.., Handbook of Physiology. Section 1: Neurophysiology,
American Physiological Society, New York, 1987, pp. 187–211.
w18x J.J. Jaeger, A.H. Lockwood, D.L. Kemmerer, R.D. Van Valin, B.W.
Murphy, H.G. Khalak, A positron emission tomographic study of
regular and irregular verb morphology in English, Language 72
Ž1996. 451–497.
w19x J.W. King, M. Kutas, A brain potential whose latency indexes the
length and frequency of words, CRL Newslett. 10 Ž1995. 1–9.
w20x M. Kutas, C. Van Petten, Event-related brain potential studies of
language, in: P.K. Ackles, J.R. Jennings, M.G.H. Coles ŽEds..,
Advances in Psychophysiology, Vol. 3, JAI Press, Greenwich, CT,
1988, pp. 129–187.
w21x M. Kutas, C. Van Petten, Psycholinguistics electrified, in: M.A.
Gernsbacher ŽEd.., Handbook of Psycholinguistics, Academic Press,
San Diego, CA, 1994, pp. 83–143.
w22x M. Kutas, R. Kluender, What is who violating: a reconsideration of
linguistic violations in light of event-related brain potentials, in:
H.-J., Heinze, T.F. Munte,
G.R. Mangun ŽEds.., Cognitive Electro¨
physiology, Birkhauser,
Boston, MA, 1994, pp. 181–210.
¨
w23x M. Kutas, S.A. Hillyard, Reading senseless sentences: brain potentials reflect semantic incongruity, Science 207 Ž1980. 203–205.
w24x M. Kutas, S.A. Hillyard, Event-related brain potentials to grammatical errors and semantic anomalies, Mem. Cogn. 11 Ž1983. 539–550.
w25x A. Laudanna, C. Burani, Distributional properties of derivational
affixes: implications for processing, in: L.B. Feldman ŽEd.., Mor-
w26x
w27x
w28x
w29x
w30x
w31x
w32x
w33x
w34x
w35x
w36x
w37x
w38x
w39x
w40x
w41x
w42x
w43x
w44x
w45x
w46x
w48x
51
phological Aspects of Language Processing, Erlbaum Press, Hillsdale, NJ, 1995, pp. 345–364.
B. MacWhinney, J. Leinbach, Implementations are not conceptualizations: revising the verb learning model, Cognition 40 Ž1991.
121–157.
L. Manelis, D. Tharp, The processing of affixed words, Mem. Cogn.
5 Ž1977. 690–695.
G.F. Marcus, U. Brinkmann, H. Clahsen, R. Wiese, S. Pinker,
German inflection: the exception that proves the rule, Cogn. Psychol. 29 Ž1995. 189–256.
R. McKinnon, L. Osterhout, Constraints on movement phenomena
in sentence processing: evidence from event related brain potentials,
Unpublished manuscript, University of Washington, Seattle, WA,
1994.
H. Meier, Deutsche Sprachstatistik, Buske, Hamburg, 1964.
T.F. Munte,
H.-J. Heinze, ERP negativities during syntactic process¨
ing of written words, in: H.-J., Heinze, T.F. Munte,
G.R. Mangun
¨
ŽEds.., Cognitive Electrophysiology, Birkhauser,
Boston, MA, 1994,
¨
pp. 211–238.
T.F. Munte,
E. Anvari, M. Matzke, S. Johannes, Brain potential
¨
correlates of umber errors in the Turkish language, Neurosci. Lett.
199 Ž1995. 57–60.
T.F. Munte,
H.-J. Heinze, G.R. Mangun, Dissociation of brain
¨
activity related to syntactic and semantic aspects of language, J.
Cogn. Neurosci. 5 Ž1993. 335–344.
T.F. Munte,
H.-J. Heinze, H. Prevedel, Erreigniskorrelierte Poten¨
tiale reflektieren semantische und syntaktische Fehler bei der
Sprachverarbeitung, Z. EEG-EMG 21 Ž1990. 75–81.
T.F. Munte,
M. Matzke, S. Johannes, Brain activity associated with
¨
number errors in words and pseudo-words, J. Cogn. Neurosci.,
Ž1997. in press.
H.J. Neville, J.L. Nicol, A. Barss, K.I. Forster, M.F. Garrett, Syntactically based sentence processing classes: evidence from event-related brain potentials, J. Cogn. Neurosci. 3 Ž1991. 151–165.
A.C. Nobre, G. McCarthy, Language related ERPs – scalp distributions and modulation by word type and semantic priming, J. Cogn.
Neurosci. 6 Ž1994. 233–255.
L. Osterhout, L. Mobley, Event-related brain potentials elicite by
failure to agree, J. Mem. Lang., 34 Ž1995. 354–375.
L. Osterhout, P.J. Holcomb, Event-related brain potentials elicited
by syntactic anomaly, J. Mem. Lang. 31 Ž1992. 785–806.
L. Osterhout, P.J. Holcomb, Event-related potentials and language
comprehension, in: M.D. Rugg, M.G.H. Coles ŽEds.., Electrophysiology of Mind: Event-Related Brain Potentials and Cognition, Oxford University Press, Oxford, 1995, pp. 175–215.
L. Osterhout, P.J. Holcomb, Event-related potentials and syntactic
anomaly: evidence of anomaly detection during the perception of
continuous speech, Lang. Cogn. Proc. 8 Ž1993. 413–437.
L. Osterhout, R. McKinnon, M. Bersick, V. Corey, On the
language-specificity of the brain response to syntactic anomalies: is
the syntactic positive shift a member of the P300 family?, J. Cogn.
Neurosci., Ž1997. in press.
R.D. Pascual-Marqui, C.M. Michel, D. Lehmann, Low resolution
electromagnetic tomography: a new method for localizing electrical
activity in the brain, Int. J. Psychophysiol. 18 Ž1994. 49–65.
M. Penke, A. Trommer, H. Weyerts, T.F. Munte,
H. Clahsen,
¨
Electrophysiological correlates of German participle formation, in:
T.F. Munte,
H.-J. Heinze, G.R. Mangun, H. Scheich ŽEds.., Map¨
ping Cognition in Time and Space, IOS Press, Amsterdam, 1997, in
press.
S.E. Petersen, P. Fox, M. Posner, M. Mintun, M. Raichle, Positron
emission tomographic studies of the processing of single words, J.
Cogn. Neurosci. 1 Ž1989. 153–170.
S.E. Petersen, P. Fox, M. Posner, M. Mintun, M. Raichle, Positron
emission tomographic studies of he cortical anatomy of single-word
processing, Nature 331 Ž1988. 585–589.
S.E. Petersen, J.A. Fiez, The processing of single words studied with
52
w49x
w50x
w51x
w52x
w53x
w54x
w55x
w56x
w57x
M. Penke et al.r CognitiÕe Brain Research 6 (1997) 37–52
positron emission tomography, Annu. Rev. Neurosci. 16 Ž1993.
509–530.
S. Pinker, A. Prince, Regular and irregular morphology and the
psychological status of rules of grammar, in: L.A. Sutton, C. Johnson, R. Shields ŽEds.., Proc. 17th Annual Meeting of the Berkeley
Linguistics Society, Berkeley Linguistics Society, Berkeley, CA,
1992.
K. Plunkett, V. Marchman, From rote learning to system building,
Cognition 38 Ž1993. 43–102.
F. Rosler,
P. Putz,
¨
¨ A.D. Friederici, A. Hahne, Event-related brain
potentials while encountering semantic and syntactic constraint violations, J. Cogn. Neurosci. 5 Ž1993. 345–362.
M.D. Rugg, Dissociation of semantic priming, word and non-word
repetition by event-related potentials, Quart. J. Exp. Psychol. 39A
Ž1987. 123–148.
D.E. Rumelhart, J.L. McClelland, On learning the past tenses of
English verbs, in: J.L. McClelland, D.E. Rumelhart, PDP Research
Group, Parallel Distributed Processing: Explorations in the Microstructure of Cognition. Vol. 2. Psychological and Biological
Models, MIT Press, Cambridge, MA, 1986, pp. 216–271.
A. Ruoff, Haufigkeitsworterbuch
gesprochener Sprache, Niemeyer,
¨
¨
Tubingen,
1981.
¨
M. Scherg, Functional imaging and localization of electromagnetic
brain activity, Brain Topogr. 5 Ž1992. 103–111.
R. Schreuder, H. Baayen, Modeling morphological processing, in:
L.B. Feldman ŽEd.., Morphological Aspects of Language Processing, Erlbaum Press, Hillsdale, NJ, 1995, pp. 131–154.
M. Taft, K. Forster, Lexical storage and retrieval of prefixed words,
J. Verb. Learn. Verb. Behav. 14 Ž1975. 638–647.
w58x M. Taft, A morphological decomposition model of lexical representation, Linguistics 26 Ž1988. 657–667.
w59x M. Taft, G. Hambly, S. Kinoshita, Visual and auditory recognition
of prefixed words, Quart. J. Exp. Psychol. 38 Ž1986. 357–366.
w60x M. Taft, Recognition of affixed words and the word-frequency
effect, Mem. Cogn. 7 Ž1979. 263–272.
w61x M. Ullman, S. Corkin, S. Pinker, M. Coppola, J.H. Growdon, J.
Locascio, Which neural structures subserve language? Evidence
from inflectional morphology, Paper presented at the 1st Annual
Meeting of the Cognitive Neuroscience Society, San Francisco, CA,
March 1994.
w62x C. Van Petten, M. Kutas, Influences of semantic and syntactic
content on open and closed class words, Mem. Cogn. 19 Ž1991.
95–112.
w63x C. Van Petten, M. Kutas, Interaction between sentence context and
word frequency in event-related brain potentials, Mem. Cogn. 18
Ž1990. 380–393.
w64x H. Weyerts, H. Clahsen, Netzwerke und symbolische Regeln im
Spracherwerb: experimentelle Ergebnisse zur Entwicklung der Flexionsmorphologie, Ling. Ber. 154 Ž1994. 430–460.
w65x H. Weyerts, T.F. Munte,
H.G.O.M. Smid, H.J. Heinze, Mental
¨
representations of morphologically complex words: an event-related
potential study with adult humans, Neurosci. Lett. 206 Ž1996.
125–128.
w66x H. Weyerts, M. Penke, U. Dohrn, H. Clahsen, T.F. Munte,
Brain
¨
potentials indicate differences between regular and irregular German
noun plurals, NeuroReport, Ž1997. in press.