See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/240751694

The time-course of inflexional morphological priming

Article in Linguistics · January 2002

DOI: 10.1515/ling.2002.011

CITATIONS READS

47 327

3 authors:

Alberto Domínguez Juan Segui

Universidad de La Laguna Paris Descartes, CPSC
69 PUBLICATIONS 719 CITATIONS 146 PUBLICATIONS 7,860 CITATIONS

SEE PROFILE SEE PROFILE

Fernando Cuetos
University of Oviedo
290 PUBLICATIONS 6,289 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

FLUIDEZ EN LECTURA Y ESCRITURA A LO LARGO DE LA VIDA. VARIABLES DEL SUJETO Y VARIABLES DE LOS ESTIMULOS (PSI2015-64174- P) View project

Visual Word Recognition; Visual Object Recognition; Age of Acquisition View project

All content following this page was uploaded by Alberto Domínguez on 04 June 2014.

The user has requested enhancement of the downloaded file.

 The time-course of inflexional
morphological priming*

ALBERTO DOMINGUEZ, JUAN SEGUI, and FERNANDO CUETOS

Abstract

In Spanish. the contrast between the suffixes -a and -o in pairs of words

such as loc-a/loc-o ‘mad woman’/‘madman’ has a morphological value
because it allows the gender selection of the stem loc. However, in pairs
such as rat-a/rat-o ‘rat’/‘moment’ this contrast has a lexical value, permit-
ting the selection of two different words. The use of these two classes of
word relationship, together with semantically related pairs such as saña–
odio ‘anger’–‘hate’, permits the study of the time course of morphological,
orthographic, and semantic priming at 32, 64, and 250 ms. SOAs. Two
experiments were carried out in which the priming effects for the previous
categories were compared with unrelated priming and identity priming in a
lexical-decision task. Morphological facilitation was obtained with 32 ms.
and 64 ms. masked presentations of the prime word and with a 250 ms.
unmasked prime. However, semantic facilitation occurred only with the
unmasked prime and orthographic facilitation occurred only for the masked
primes. Morphological, orthographic, and identity priming effects were of
comparable magnitudes for the 32 ms. masked primes, whereas for the
64 ms. masked priming, orthographic priming was half the magnitude of
morphological and identity priming. To what extent the facilitation for
morphological pairs is due to a morphological stage of processing or, rather,
to the summed effects of orthographic and semantic features is discussed.
Agreements and differences with some recent studies in other languages are
considered (Drews and Zwitserlood 1995; Frost et al. 1997; Rastle et al.
2000).

Some of the most important word-recognition models support the

requirement of a morphological treatment of the input at some level of
processing (see McQueen and Cutler 1998 for a review). The full parsing
models defend a mandatory prelexical segmentation of the word into its

Linguistics 40–2 (2002), 235–259 0024–3949/02/0040–00235

© Walter de Gruyter
236 A. Dominguez, J. Segui, and F. Cuetos

stem and affixes and lexical access of morphological relatives across the
same entry (Taft and Forster 1975). On the other hand, some full listing
models locate morphological processing at the level of the central lexicon.
All morphologically complex words are represented but the morphologi-
cal relatives’ nodes are related by facilitatory links (Colé et al. 1989;
Drews and Zwitserlood 1995; Grainger et al. 1991).
However, some other theoretical approaches to visual word recognition
do not give an explicit representation to the morphological constituents
of the word. The interactive activation model, for example (McClelland
and Rumelhart 1981), proposes letters as the unique unit with representa-
tion at prelexical levels. Also, the orthographic redundancv hypothesis
(Seidenberg 1987) defends the non-necessity of a supra-letter access code
to word recognition and therefore the uselessness of syllables or mor-
pheme units. Orthographic and semantic features occur concurrently with
morphological relations between words. Failing and failure, for example,
have in common not only some of their orthographic and phonological
characteristics but also some of their semantic features. Therefore, accord-
ing to these theories, morphology is a linguistic category that does not
have an independent entity in the access, representation, or organization
of the lexicon. Processing of the orthographic and semantic features of
words is sufficient for efficient lexical representation.
This argument has been refuted by a line of research that has attempted
to dissociate the sources of processing for orthographic, morphological,
and semantic relations between words (Beauvillain and Segui 1992;
Feldman and Andjelkovic 1992; Stolz and Feldman 1995 for review).
The priming paradigm allows manipulation of the influence that a prime
has on the processing of a target word. In general, if the relation between
prime and target is morphological, then target processing will be facili-
tated (Grainger et al. 1991). This morphological facilitation occurs
whether the prime and target are separated by a large number of other
stimuli (Feldman 1994; Feldman and Andjelkovic 1992; Fowler et al.
1985) or are presented contiguously and in the contiguous case, whether
the prime is masked or unmasked (Drews and Zwitserlood 1995; Forster
1987; Grainger et al. 1991).
Unlike morphological priming, formal priming (e.g. prime and target
sharing letters or phonemes) is dependent on the relative frequencies of
prime and target, the number of letters, and the time of exposition. With
unmasked contiguous prime–target presentation, inhibition is obtained
when a target is primed by a lower frequency orthographic neighbor,
whereas no priming is achieved at long lags (Feldman and Moskovljevic
1987; Napps 1989). When a masked prime immediately precedes the
 Inflexional morphological priming 237

target, a nonsignificant orthographic facilitation effect occurs (Segui and

Grainger 1990).
Important evidence concerning the differences between the morphology
and orthography time-course has been reported by Drews and
Zwitserlood (1995), using masked and unmasked primes in lexical-
decision and naming tasks in Dutch and German. For both masked
and unmasked primes, the morphologically related pairs (kersen–KERS
‘cherries’–‘cherry’) yielded facilitation in both tasks, whereas the ortho-
graphically related pairs (kerst–KERS ‘Christmas’–‘cherry’) yielded inhi-
bition for lexical decisions and facilitation for naming. These results agree
with those of Segui and Grainger (1990), who obtained inhibition from
an orthographically related masked prime.
In contrast with inflectional morphology, derivational morphology
allows the manipulation of semantic relations between morphological
relatives. Derived words may maintain a close or a distant semantic
relation. This peculiarity has been exploited in order to find the degree
of relation between morphology and meaning. For example Bentin and
Feldman (1990) compared morphological priming effects for semantically
related and unrelated pairs as a function of lag. Whereas semantic facilita-
tion was significant at 0 lag and disappeared at a lag of 15, the morpholog-
ical, semantically unrelated priming condition exhibited a small but
significant facilitation at lag of 0 and 15. The morphological, semantically
related priming produced facilitation similar to the semantically associ-
ated pairs at 0 lag, and this facilitation decreased to equal that of
the morphological, semantically unrelated priming condition at lag 15.
Also Napps (1989), using contiguously presented visual primes and
targets, found facilitation for morphologically related words but not for
semantically or orthographically related words.
The results reported support some differences in the time-course and
tendency of priming for three types of relation: morphological, ortho-
graphic, and semantic. The aim of the present research is to study
systematically these three types of effect, using both masked and
unmasked primes in a lexical-decision task with three different prime–
target stimulus onset asynchronies (SOAs) and under well-controlled
conditions.
Marslen-Wilson et al. (1994) emphasized the importance of providing
systematic accounts of morphological representation and access for indi-
vidual languages because the morphological characteristics of languages
differ. Gender, for example, is represented by suffixes in the inflected
morphology of Spanish nouns and adjectives but this is not the case with
English morphology. However, Spanish shares with English and many
other Indo-European languages a concatenative morphology in which
238 A. Dominguez, J. Segui, and F. Cuetos

the different morphemes, roots, suffixes, and prefixes are combined in a

linear fashion and contrasts with Hebrew and other semitic languages
that have a nonlinear morphology (Frost et al. 1997). These differences
allow comparative research into the influence of the superficial features
of each language on a supposedly universal mechanism of representation
and morphological processing. Spanish morphology is very rich and
complex. Gender, in particular, is, with number, the basic category of
inflexional morphology. Words denoting biological and grammatical
gender add to their root morpheme the marker -a, to make the feminine
gender (e.g. loc-a ‘madwoman’), and -o, to obtain the masculine gender
(e.g. loc-o ‘madman’). However, for some nouns and adjectives, the
application of this ‘‘gender rule’’ produces a morphologically unrelated
but orthographically related word. This is the case with the noun rata, a
feminine Spanish word to designate ‘rat’. If the final letter -a is replaced
by -o, the word rato is obtained, which does not mean ‘male of the rat’,
but ‘a moment’. The word rato is orthographically but not morphologi-
cally or semantically related to the word rata. Therefore, the contrast
between the suffixes -a/-o in Spanish has a lexical value for rata/rato
pairs, because it permits the selection of two different words. But this
contrast has no lexical value for loca/loco pairs, because in this case, the
suffixes -a/-o allow the gender selection of a stem (loc-). At representation
level, rata and rato require independent lexical entries, whereas loco and
loca may share the same lexical entry, the stem loc-, according to some
morphological models (Taft 1979, 1994; Taft and Forster 1975).
Thus, Spanish permits the selection of morphologically related pairs
(e.g. LOCA–loco), orthographically related pairs (e.g. RATA–rato), and
semantically related pairs (e.g. luto/pena ‘mourning’–‘sadness’) of words
that are equal in length within and between pairs and have the same
number of shared letters. This constraint is important as it allows us to
control three aspects. First, the length of words can be reduced to four
or five letters, making our stimuli comparable to those used to study
formal priming (orthographic or phonological ). This is relevant because
Forster et al. (1987) found facilitation with seven- or eight-letter words
(bontrast–CONTRAST ), whereas Segui and Grainger (1990) obtained
inhibition with four-letter words (blur–blue) Second, the prime length in
our experiments can be equaled to the target length, as can the position
of the shared letters. Stolz and Feldman (1995) provide some evidence
of a different result for morphological priming at short lags when prime
and target are similar or not similar in length. Similar length in prime and
target yielded facilitation whereas different lengths reversed this tendency
to inhibition. Third, because morphologically related primes and targets
in studies in English, German, Dutch, French, or Serbo-Croatian do not
 Inflexional morphological priming 239

always have the same number of letters, it is problematic to compare

these with identity priming in which, by definition, the same word is
repeated.
Finally, it is important to mention the type of morphologically related
words that we are going to use: inflected words. Inflectional morphology
(gender suffixes) has special properties with respect to derivational mor-
phology (Sandra 1994). The function of the inflected suffix is basically
syntactic, showing how a word (verb, determiner, adjective, noun, or
pronoun) maintains a relation to another constituent of the sentence (e.g.
agreement). The inflections are limited in number and are rule-governed.
Also, words become inflected by addition of a suffix whereas derivational
systems use suffixes and prefixes to obtain morphologically complex
words. With these properties, Sandra defends a system of representation
for inflected suffixes separate from the lexicon where content-derived or
nonderived words are stored. Derived and compound words, in contrast,
often change the semantic, syntactic, and phonological properties of the
roots and are very much greater in number than inflections. These varia-
tions in transparency could produce important differences in the access
or representation for inflected or derived words. Of course, irregularity
and orthographic opacity may be found in inflected forms, such as
irregular past tense verbs in English or Spanish. The dual-mechanism
account (Pinker 1991; Prasada and Pinker 1993) defends a rule-based
mechanism for regular words and an associative mechanism for irregular
inflected words.
In our experiments, the stimuli are gender-inflected Spanish words that
could be obtained by the rule ‘‘attach to the end of the root an -a if
feminine or an -o if masculine.’’ This regularity and transparency favors
a possible segmentation procedure to accede the specific word that is
expected to emerge in morphological priming but not in orthographic or
semantic priming.

Experiment I

The prime–target relations were varied across three different SOAs: 32,
64 (with masked prime presentation), and 250 ms. (with unmasked prime
presentation). Because masked and unmasked primes at SOAs shorter
than 300 ms. presumably preclude expectancy-based priming effects
(see Neely 1991 for a review in the semantic-priming domain), whereas
unmasked primes at longer SOAs do not, we used SOAs less than 250 ms.
to determine the degree to which morphological priming effects are
‘‘automatic.’’
240 A. Dominguez, J. Segui, and F. Cuetos

Based on prior research in other languages, we expected orthographic,

phonological, and morphological priming to occur very early. For exam-
ple, Ferrand and Grainger (1992), Lukatela and Turvey (1990), and
Perfetti et al. (1988) have demonstrated that orthographic and phonologi-
cal information becomes available after 33 ms. Also, Perfetti and Tan
(1998) obtained graphic facilitation at 43 ms. and inhibition at longer
SOAs. The phonological facilitation was obtained later (57 ms.) and the
semantic facilitation began to appear at 85 ms. SOA for precise meaning
relations and at 115 ms. for vague meaning relations. Therefore, if the
morphological structure of the word is computed later than the processing
of phonology and orthography, the 32 and 64 ms. masked priming condi-
tions might reflect this difference. The 64 ms. SOA is the interval systemat-
ically used by researchers with masked priming of morphologically related
words (Drews and Zwitserlood 1995; Forster et al. 1987; Grainger et al.
1991). On the other hand, with an unmasked prime at a 250 ms. SOA
the prime may be consciously identified. Due to the fact that unmasked
orthographic priming may produce inhibition (Segui and Grainger 1990;
see also Colombo 1986 for phonological priming and Dominguez et al.
1997 for syllabic priming), if morphological priming is facilitatory under
conditions in which orthographic priming is inhibitory, then this would
be evidence that morphological priming produces, at this SOA, a course
of processing different from orthographic priming. The probability of
finding semantic priming with associate words decreases as a function
of SOA (see Lukatela and Turvey 1994a; Perea et al. 1995; Perea and
Gotor 1997 for alternative points of view). Thus, morphological facil-
itation should be preserved in masked priming, whereas semantic
facilitation will not appear. However, it is predicted that the semantic
facilitation should appear at the 250 ms. SOA.

Method

Participants. One hundred and eighteen psychology undergraduates at

Oviedo University, in Spain, participated in the experiment for course
credit (40 students in the 32 ms. SOA condition, 38 in the 64 ms. SOA,
and 40 in the 250 ms. SOA conditions). All were Spanish native speakers
with normal or corrected-to-normal vision.

Stimuli. Forty-eight target stimuli were each paired with a related and
an unrelated word. In the morphological category, the target loco
‘madman’ was paired with LOCA ‘madwoman’ for the related condition
and with CERA ‘wax’ for the unrelated condition. Rato ‘moment’ was
 Inflexional morphological priming 241

paired with RATA ‘rat’ in the orthographic condition and with PERA
‘pear’ in the corresponding unrelated condition. Finally pena ‘sadness’
was paired with LUTO ‘mourning’ in the semantic condition and with
LOBO ‘wolf ’ in the unrelated pair. Both orthographically and morpho-
logically related pairs were composed of primes and targets sharing all
letters except the final letter. The semantic pairs differed in all letters.
Half of the experimental primes finished in -a when the target finished
in -o and vice versa. The semantically related pairs were obtained from
normative studies.1
Sixteen pairs of each type were generated. Eight were presented in the
related conditions and eight in the unrelated condition for half of the
participants and this was reversed for the other half. Each subject saw
each prime and target once during the experiment. The prime words had
a lower lexical printed frequency than the target. This frequency relation
was constrained target by target (see Table 1).
A list of fillers, word–word and word–nonword pairs, was introduced
to reduce the percentage of orthographically related pairs (12.5%) and
semantically related pairs (8.30%). The number of related and unrelated
pairs was the same for word and nonword targets. All experimental
stimuli were four or five letters long, and they ended in -a when the target
ended in -o and vice versa. The inflection -o corresponds to masculine
and -a to feminine gender. These characteristics were distributed half and
half across each category. Some fillers ended with different inflectional
suffixes (-e, -al, etc.) and they were of four, five, six, or seven letters to
approximate characteristics of the stimuli to normal Spanish distribution.
All stimuli were Spanish nouns or adjectives. Each subject received 192
pairs of stimuli, 96 word and 96 nonword targets; 48 of the word pairs

Table 1. Mean printed frequency (F) and standard deviation (S.D.) for targets and primes
in each of the experimental categories

Type of priming Target Related prime Unrelated prime

F S.D. F S.D. F S.D.

Experiment 1
Morphological 80.68 73.16 36.09 36.42 26.50 27.53
Orthographic 84.25 57.30 11.50 13.72 10.72 15.40
Semantic 82.34 84.50 12.68 7.21 11.76 6.08

Experiment 2
Morphological 60.07 64.67 27.00 33.46 22.55 31.25
Orthographic 65.60 54.32 10.42 11.75 9.38 12.82
242 A. Dominguez, J. Segui, and F. Cuetos

were experimental word–word pairs (24 related and 24 unrelated, eight

of each type) and 48 were unrelated filler pairs.

Design. A 3×3×2 design was generated. The first factor was called
SOA and included these three levels: 32, 64, and 250 ms. The second
factor was type of priming, and its three levels were morphological,
orthographic, and semantic. The third factor, relation, included two
levels: related and unrelated.

Procedure. The stimuli were presented in the center of the screen of a

486 PC with a 70 Hz refresh rate. The letters, in Courier font, appeared
as white characters on a dark background. The primes were presented in
uppercase, while the targets were in lowercase. Each character covered
approximately 0.38◊ of visual angle from a distance of 60 cm. These
display characteristics were the same in all experiments. The masking
procedure was in general the same used by Forster and Davis (1984) and
by Grainger et al. (1991). The sequence of events that occurred in each
trial was, first, a mask of hash marks (#) that exactly covered the number
of spaces of the prime stimulus and was presented for 500 ms. Second,
the prime was exposed in the same place for 32 ms. or 64 ms. and
immediately followed by the target stimulus, which remained on the
screen until subject response. For the 250 ms. SOA a point of fixation
appeared in the center of the screen and remained there for one second,
being immediately replaced by the prime word, which was exposed for
200 milliseconds. There was, then, a 50 ms. blank interval, and then the
target string appeared and remained there until subject response. At six
different moments, a message on the screen asked the subjects to report
the last pair of stimuli seen. Thus, it was possible to ensure that the
subject was paying attention to the prime. Participants in the masked
conditions were not informed of the prime presence. They were told that
one stimulus would appear in the center of the screen and their task
consisted of pressing a key as rapidly and as accurately as possible,
‘‘YES’’ if a word and ‘‘NO’’ if a nonword appeared. The computer
recorded response keys and latencies.

Results

Mean reaction times (RTs) for correct responses in each condition for
32, 64, and 25 ms. SOAs and the mean percentage of errors are presented
in Table 2. ANOVAs were carried out by subjects (F1) and by items
(F2), taking as independent factors the SOA, the type of priming, and
 Inflexional morphological priming 243

Table 2. Mean reaction times (in milliseconds) and error percentages (%) in experiment 1
(32, 64, and 250 ms. SOA)

Type of priming Related Unrelated Effect % Fac.

Subj. Item

32 ms.
Morphological 648 677 +29* 73 68
(LOCA–loco) (1.87) (0.94) (−0.93)
Orthographic 656 683 +27* 63 69
(RATA–rato) (1.87) (2.19) (+0.32)
Semantic 691 688 −3 50 50
(LUTO–pena) (1.87) (3.12) (+1.25)

64 ms.
Morphological 618 686 +68* 84 94
(LOCA–loco) (4.60) (3.28) (−1.32)
Orthographic 646 674 +28* 63 69
(RATA–rato) (2.63) (2.30) (−0.33)
Semantic 687 701 +14 53 63
(LUTO–pena) (2.96) (3.61) (+0.65)

250 ms.
Morphological 637 690 +53* 80 88
(LOCA–loco) (0.60) (2.81) (+2.21)
Orthographic 722 710 −12 40 25
(RATA–rato) (7.18) (5.00) (−2.18)
Semantic 667 712 +45* 83 82
(LUTO–pena) (3.12) (2.81) (−0.31)

The Effect column is the result of subtracting related from unrelated means in each of the
three types of priming. The percentages of subjects and items that produce facilitation
(% Fac.) for each of the effects are presented. Probabilities less than 0.05 (*) and less than
0.01 (**) are indicated here and in the following tables.

the relation. In the F1 ANOVAs the SOA factor was taken between-
subjects while the type of priming and the relation were taken within-
subjects. On the contrary, in F2 ANOVAs, the type of prime factor was
between-items and the prime duration and the relation were considered
within-items. Latencies exceeding 1200 ms. or not reaching 200 ms. were
excluded from the analysis (4.27% of the data). These two cutoff points
were the same in the next experiment.
Table 3 shows the overall analysis and the three separate analyses for
each 32, 64, and 250 ms. SOA. The relation effect showed significant
facilitation produced by the related prime–target words on the unrelated
pairs. The three-way significant interaction denotes the change of priming
244 A. Dominguez, J. Segui, and F. Cuetos

Table 3. Analyses of variance with the stimulus onset (SOA), type of priming (TP), and
relation (R) as variables for exeriment 1

Source F1 F2
Df F Mse Df F Mse

Overall analysis
SOA 2,115 0.82 34900.7 2,90 6.02** 1243.64
TP 2,230 27.14** 2334.86 2,45 3.87* 7104.92
R 1,115 57.47** 2347.61 1,45 48.81** 1014.37
SOA×TP 4,230 10.02** 2334.86 4,90 8.01** 1243.64
SOA×R 2,115 2.37 2495.31 2,90 1.77 1175.50
TP×R 2,230 9.62** 2296.99 2,45 7.07** 1014.37
SOA×TP×R 4,230 5.36** 2296.99 4,90 4.63** 1175.50

32 ms. SOA
TP 2,78 9.68** 2689.70 2,45 2.55 2689.70
R 1,39 7.28** 2477.39 1,45 8.56** 915.35
TP×R 2,78 3.46* 1761.99 2,45 2.59 915.35
M 1,39 11.85* 1361.48 1,15 5.13* 1226.46
O 1,39 6.19* 2250.92 1,15 7.87* 796.00
S 1,39 0.60 2388.96 1,15 0.04 723.59

64 ms. SOA
TP 2,74 17.03** 2225.03 2,45 4.35* 3255.94
R 1,37 30.54** 2514.21 1,45 28.18** 1125.45
TP×R 2,74 5.80** 2551.36 2,45 4.48* 1125.45
M 1,37 35.33** 2476.95 1,15 103.0** 319.39
O 1,37 5.08* 2984.53 1,15 5.08* 2014.38
S 1,37 1.73 2155.46 1,15 1.78 1042.57
M–O 1,37 5.67* 2631.95 1,30 4.07* 1166.89

250 ms. SOA

TP 2,78 17.80** 3178.75 2,45 7.42** 3433.09
R 1,39 24.18** 2059.78 1,45 10.64** 1329.03
TP×R 2,78 9.87** 2590.67 2,45 8.02** 1329.03
M 1,39 20.87** 2704.98 1,15 14.31** 1151.19
O 1,39 0.92 3269.20 1,15 1.49 1717.99
S 1,39 32.75** 1266.95 1,15 14.69** 1117.91

Morphological (M ), orthographic (O), and semantic effects on the unrelated pairs are also
provided. M–O is the difference between the morphological and the orthographic
facilitation.

pattern in the orthographic and semantic type of relation across the three
SOAs, whereas morphological priming presented persistent facilitation.
The separate analyses (see Table 3) show significant facilitation for the
orthographic and morphological related pairs at 32 and 64 ms. and a
nonsignificant effect of the semantically related words. However, the
 Inflexional morphological priming 245

morphological facilitation at 64 ms. prime duration was significantly

greater (68 ms.) than the orthographic facilitation (28 ms.) (see Table 3).
On the other hand, the 250 ms. SOA analysis yielded significant facilita-
tion for morphologically and semantically related pairs and nonsignificant
inhibition for orthographically related pairs.

Discussion

The orthographic facilitation effect of 27 ms. at 32 ms. SOA supports an

early and automatic orthographic/phonological activation during the first
few milliseconds of stimulus processing. The similar morphological facili-
tation of 29 ms. can be interpreted, moreover, as the product of the
orthographic activation. At this SOA the semantic relations do not seem
to induce any type of priming. Similar orthographic or phonological
facilitation has been found in other priming studies at a similar SOA for
lexical-decision (Perea and Algarabel 1992), naming (Lukatela and
Turvey 1994b), and word-identification tasks (Perfetti et al. 1988). At
64 ms. SOA, however, there was longer facilitation for morphologically
related pairs than for orthographically related pairs. When the interval
of influence of the prime stimuli on the target is longer, the shared root
introduces greater facilitation (68 ms. vs. 28 ms.). This activation cannot
be reduced to a purely orthographic effect as occurred at a shorter SOA,
nor to a semantic effect because no significant facilitation was found.
Garcı́a-Albea et al. (1998), with a very similar manipulation and also
using Spanish materials, found significant facilitation at 64 ms. masked
priming for morphological pairs but nonsignificant facilitation for ortho-
graphically related pairs, although they did not directly compare the two.
At 250 ms., on the contrary, the orthographic facilitation disappears
and a tendency to inhibition emerges as a result of the competition of
the previously activated prime with the target word that is going to be
recognized (Segui and Grainger 1990). This is the usual pattern of ortho-
graphic priming when the prime is lower in frequency than the target.
The absence of inhibition for morphologically related pairs when they
maintain exactly the same orthographic relation as the orthographically
related pairs points to a different processing procedure for morphological
relations. The fact that at 64 ms. SOA the morphological facilitation was
greater than at 250 ms. SOA can be explained because at the longer SOA
the orthographic relation diminishes, reducing the morphological facilita-
tion for morphologically related words. A similar result and explanation
can be found in the study of Grainger et al. (1991). On the other hand,
the semantically related pairs now began to yield significant facilitation
246 A. Dominguez, J. Segui, and F. Cuetos

over the unrelated pairs. Some other studies have demonstrated a dissoci-
ation between morphological and semantic priming, with unmasked con-
tiguous presentation (Napps 1989) or manipulating the number of
merged lags (Bentin and Feldman 1990). Our experiment permits a
picture of the time-courses for the three types of relation explored, but
an issue that remains to be clarified is the nature of the facilitation
obtained at short and masked prime presentation. We know that the
facilitation at 6 ms. with morphological pairs is different from the facilita-
tion obtained with orthographic pairs, but we do not know whether the
significant facilitation with morphological pairs at 32 is also morphologi-
cal or could be considered as a more superficial orthographic activation.

Experiment 2

The aim of this experiment is to compare the magnitude of the morpho-

logical priming effect with respect to repetition priming. Traditionally,
an analogous amount of morphological priming compared with repetition
priming (full priming) suggests that morphologically related prime and
target share a lexical entry (Forster et al. 1987). If full priming is obtained
only for morphological pairs at 32 ms. SOA and not for orthographic
pairs, we could conclude that the nature of the morphological activation
at this SOA is now different from the orthographic activation. Indeed,
we hope that at 64 ms. SOA full priming will only be obtained in the
case of morphological pairs, supporting the differences between ortho-
graphic and morphological priming at this SOA found in experiment 1.

Method

Subjects. Ninety second-year undergraduates at La Laguna University,

Spain, participated in this experiment (48 subjects took part in the 32 ms.
SOA and 42 participated in the 64 ms. SOA) for course credit. None of
them had participated in the previous experiment. All were Spanish native
speakers with normal or corrected-to-normal vision.

Stimuli and design. A set of morphological stimuli was composed.

Twenty-four target words were paired with three types of prime: first,
the same word as the target (identity: LOCO–loco), second, a morpholog-
ically related word (related: LOCA–loco), and third, a prime word with-
out orthographic, semantic, or any other relation to the target (unrelated:
CERA–loco). Another set of orthographic stimuli was composed of
 Inflexional morphological priming 247

twenty-four word targets that were paired with three types of prime;
repetition: RATO–rato, orthographically (but not morphologically)
related: RATA–rato, and unrelated: PERA–rato. A 2×2×3 design was
used: SOA (32 and 64 ms.) by type of priming (morphological and
orthographic pairs), by relation (repetition, related, unrelated pairs). The
mean frequency of stimuli for each category can be seen in Table 1.

Procedure. Basically the same procedure was followed as that of the

first experiment. The time of exposition of the prime stimulus was 32
and 64 ms. Each subject received eight pairs of stimuli by category, 48
experimental pairs in all.

Results

Table 4 shows the mean lexical decision times and percentage of errors
(between brackets) for experiment 2. Latencies (excluding errors) and
errors were submitted to ANOVAs (F1 and F2 subjects and items analy-
ses) where the principal factors were SOA, type of prime, and relation.
Reaction times exceeding 1200 ms. or not reaching 200 ms. were removed
from the analyses (2.17% of the data).
An overall analysis was carried out including the three factors, and
two other separate analyses were carried out for each one of the SOA
levels (32 and 64 ms.). Table 5 shows these three analyses of variance.
The overall analysis showed a significant effect of the relation factor
due to the facilitation from the identity pairs (53 ms.) and the related
pairs (44 ms.) on the unrelated pairs. None of the factors interacted with
each other. Nevertheless, separate analyses were carried out for each
prime duration level. A particular point of interest was that at 32 ms.
prime duration, the orthographic facilitation was similar to the identity
facilitation (2 ms. of difference), as was the morphological facilitation
with respect to the identity priming (3 ms. of difference) (see Table 4).
At this SOA we obtained full priming for morphologically and ortho-
graphically related pairs. However, at 64 ms. prime duration, the ortho-
graphic facilitation was significantly smaller than the identity facilitation
(21 ms. of difference), whereas the morphological pairs continued to
produce full priming (9 ms. of difference).

Discussion

The fact that the orthographic pairs produced facilitation as great as the
identity pairs at 32 ms. SOA supports an extended spreading activation
Table 4. Mean reaction times (in ms.) and error percentages (%) in experiment 2, introducing an identity priming, with two different SOAs (32 ms.
and 64 ms.)

Priming Priming condition Priming effect

Identity Related Unrelated Id. N.Id. Id.–Rel % Fac.
Subj. Item

32 ms. SOA
Morphological 633 636 683 +50* +47* −3 75 73
(LOCA–loco) (1.82) (5.46) (3.38) (+156) (−2.08) (−3.64)
Orthographic 651 653 693 +42* +40* −2 75 69
(RATA–rato) (3.38) (1.82) (2.60) (−0.78) (+0.78) (+1.56)
248 A. Dominguez, J. Segui, and F. Cuetos

64 ms. SOA
Morphological 618 627 682 +64* +55* −9 83 83
(LOCA–loco) (3.27) (2.98) (4.46) (+1.19) (+1.48) (+0.29)
Orthographic 637 658 693 +56* +35* −21* 75 62
(RATA–rato) (2.67) (2.98) (2.08) (−0.59) (−0.90) (−0.31)

The identity effect (Id.), the non–identity effect (N.Id.), and the result of subtracting both (Id–Nid) are shown. The percentages of subjects and
items that produce facilitation (% Fac.) are shown.
 Inflexional morphological priming 249

Table 5. Overall and separate analyses of variance with stimulus onset asynchrony (SOA),
type of priming (TP), and relation (R) as variables for the two prime durations manipulated

Source F1 F2
Df F Mse Df F Mse

Overall analysis
SOA 1,88 0.20 21827.6 1,46 1.17 1069.53
TP 1,88 19.69** 2086.68 1,46 1.87 10441.8
R 2,176 66.91** 2142.41 2,92 52.25** 1531.60
SOA×TP 1,88 0.42 2086.68 1,46 0.18 1069.53
SOA×R 2,176 1.18 2142.41 2,92 0.99 1324.27
TP×R 2,176 0.82 2309.16 2,92 0.92 1531.60
SOA×TP×R 2,176 0.23 2309.16 2,92 0.29 1324.27

32 ms. SOA
TP 1,47 7.95** 2016.39 1,46 1.73 4591.37
R 2,94 33.29** 1924.32 2,94 22.19** 1490.49
RTP×R 2,94 0.16 2399.91 2,92 0.16 1490.49
Morphological set analyses
R 2,94 15.82** 2355.46 2,46 10.17** 1913.98
I 1,47 29.88** 1970.02 1,23 19.15** 1613.62
N.Id. 1,47 17.72** 2975.65 1,23 18.29** 1498.91
I–M 1,47 0.08 2120.71 1,23 0.04 2629.41
Orthographic set analyses
R 2,94 13.82** 968.77 2,46 12.98** 1067.00
I 1,47 24.22** 1770.23 1,23 16.48** 1346.13
N.Id. 1,47 21.43** 1801.01 1,23 16.25** 1183.76
I–O 1,47 0.05 2335.07 1,23 0.16 671.11

64 ms. SOA
TP 1,41 11.68** 2167.68 1,46 1.71 6919.96
R 2,82 33.74** 2392.42 2,92 35.35** 1365.37
TP×R 2,82 0.88 2205.13 2,92 1.14 1365.47
Morphological set analyses
R 2,82 25.83** 1914.83 2,46 19.33** 1603.04
I 1,41 53.05** 1594.84 1,23 29.50** 1769.76
N.Id. 1,41 25.51** 2429.75 1,23 27.77** 1437.20
I–M 1,41 1.02 7874.30 1,23 0.51 1602.15
Orthographic set analyses
R 2,82 12.38** 2682.72 2,46 16.71** 1127.71
I 1,41 23.73** 2732.39 1,23 29.14** 1264.82
N.Id. 1,41 8.14** 3202.53 1,23 11.10* 1317.78
I–O 1,41 4.11* 2113.24 1,23 6.30* 800.52

Identity (I ) and non–identity (N.Id) effects and the difference between the Identity condition
and the Related condition (I–M or I–O) for each set of stimuli are shown.
250 A. Dominguez, J. Segui, and F. Cuetos

from an orthographic description of the input stimulus to all orthographi-

cally related word units or, in terms of access, contact with the same
entry in the lexicon. We maintain that, at this SOA, there is not a special
treatment for morphological pairs in relation to orthographic pairs at
this prime duration. In contrast, at 64 ms. SOA the morphologicalty
related pairs produced full priming (9 ms. of difference with the identity
condition) whereas orthographic pairs do not (21 ms. of difference with
the identity condition). The ‘‘full priming’’ obtained for morphologicaly
related pairs supports the assumption that both words, prime and target,
could reach the same lexical entry. This is the hypothesis that has been
proposed by Forster et al. (1987). However, an alternative and compatible
point of view of the results could be presented. An interactive activation
system that represents at the word level separate entries for loco and loca
interconnected with facilitation links may produce full priming for mor-
phological pairs (Drews and Zwitserlood 1995; Grainger et al. 1991).
Even a model that does not represent morphological processing at any
level of processing (Seidenberg 1987), only the summing up of semantic
and orthographic activation, could account for these results.

General discussion

A systematic comparison of the time course for morphological, ortho-

graphic, and semantic relations was conducted across three SOAs (e.g.
32, 64, and 250 ms.). The objectives were, first, to take advantage of
Spanish inflexional morphology, which allows rigorous control of some
of the variables that may influence RTs, such as the perfect matching in
length between prime and target and the percentage of shared letters,
to establish consistent contrasts between priming categories; second, to
provide data supporting a specific stage of morphological processing,
nonreducible to the sum of orthographic plus semantic activation.
The results of two experiments show a particular evolution for each
priming relationship (Figure 1). The morphologically related words pro-
duced consistent facilitation during the three explored SOAs, 32, 64, and
250 ms. This facilitation reaches a maximum value at 64 ms. SOA. This
prime duration was sufficient to compute the morphological attributes of
the word. The orthographically related words produced about the same
facilitation at 32 ms. and at 64 ms. SOA. This facilitation reverses to
nonsignificant inhibition at 25 ms. SOA. The morphological facilitation
at 64 ms. SOA reaches 68 ms., twice the orthographic facilitation. Finally,
the semantically related pairs produced significant facilitation only at
250 ms. SOA and null effects at shorter SOAs. This time course suggests
 Inflexional morphological priming 251

Figure 1. Priming at different SOAs in experiment 1: the positive effects represent facilitation
whereas the negative effects represents inhibition over the unrelated pairs

a different temporal influence for each lexical attribute: first, very early
computation of both the orthographic and phonological properties
(which, because of the characteristic transparence of Spanish orthogra-
phy, are not dissociated in our materials); second, morphological process-
ing, and later, computation of the semantic attributes (see Figure 1).
Experiment 2 replicated the temporal pattern of experiment 1. The
orthographic and the morphological pairs yielded full priming at 32 ms.
of prime exposition, whereas only the morphological pairs showed full
priming at 64 ms. prime duration. Whereas the orthographic priming is
analogous to repetition priming at 32 ms. SOA, the equivalence for
morphological priming is subsequent (64 ms. SOA).
These results replicate the data of a new and very relevant morphologi-
cal priming study that was published after our article had been written.
Rastle et al. (2000) contrasted morphological (derivational ), ortho-
graphic, and semantic priming along three SOA conditions (43 ms.,
72 ms., and 230 ms.). They found consistent priming for semantically
and morphologically related forms across all SOAs. These effects were
significantly greater than those produced by orthographic or semantic
priming and not different from those of identity facilitation. More impor-
tantly, in a second experiment, they contrasted morphological priming
(adapter–adaptable) with orthographic plus semantic priming without
morphological relation (screech–scream) to observe whether or not mor-
phological priming could be reducible to the sum of semantic and ortho-
252 A. Dominguez, J. Segui, and F. Cuetos

graphic activation. The results showed that semantic and orthographic

overlap was not sufficient to produce facilitation at the shorter SOAs,
whereas the morphological relation continued to produce this.
From our point of view, Rastle et al.’s results could be considered as
complementary of our results (and vice versa) in consideration of two
points: first, Rastle and colleagues use derivational morphology, whereas
our morphological priming uses gender-inflected words; and second, as
a consequence, (a) our control of orthographic overlapping is better than
Rastle et al.’s, (b) they have better control of semantic relationships, and
(c) they included a critical category of morphologically related words
that were semantically unrelated. A comparative interpretation of our
morphological priming data with Rastle et al.’s experiments and various
other relevant studies may be pertinent in two aspects: orthographic
contrast and semantic contrast.

Orthographic contrast

The stimuli of our morphologically related category (loca–loco) and

orthographically related category (rata–rato) are perfectly matched in
length and differed in only one letter in prime and target: they could be
defined as orthographic neighbors. The use of four- and five-letter words
and control of the relative prime–target frequency permitted a direct
comparison of the results with those of the literature on orthographic
processing. As in our two experiments, what has normally been found
with masked priming is facilitation at short SOAs and with unmasked
priming, inhibition at long SOAs (Colombo 1986; Grainger 1990; Segui
and Grainger 1990; Perfetti and Tan 1998). But it is important to point
out that our second experiment showed a similar facilitation for ortho-
graphic, morphological, and identity pairs at shorter SOA (32 ms.), indi-
cating that our orthographic manipulation is particularly appropriate.
This panorama changes for orthographic conditions (formal priming)
in some studies centered on morphological processing. The lack of rigor-
ous control of orthographic variables in these studies determined that
orthographic priming did not reach significant facilitation. For example,
Drews and Zwitserlood (1995), comparing orthographic and morphologi-
cal priming, found orthographic inhibition irrespective of the prime dura-
tion and the masked or unmasked presentation. Also, Rastle et al. (2000)
found nonsignificant facilitation (of 19 and 15 ms.) at 43 and 72 ms.
SOAs. Drews and Zwitserlood (1995) used target words shorter than
(and completely included in) the prime and with equal frequency for
primes and targets (i.e. kerst–KERS ‘Christmas’–‘cherry’). This circum-
 Inflexional morphological priming 253

stance may favor a very early activation and inhibition of the target
word. On the other hand, Rastle et al., with a similar orthographic
overlapping (departure–depart), obtained a tendency to facilitation. This
change in the tendency with respect to the article of Drews and
Zwitserlood could be due to the frequency relation between prime and
target, not reported by Rastle et al. (see Segui and Grainger 1990 for a
direct manipulation of the prime–target relation frequency).
Also, Frost et al. (1997) conducted a masked priming experiment in
which the prime shared with the target some letters that conformed to
what in Hebrew is known as the word pattern in opposition to the root
pattern. This is, in fact, formal priming. The result failed to obtain
orthographic facilitation at 43 ms. SOA, also in contrast with our results.
Once again, their primes share less than 50% of the letters with their
targets, whereas our materials overlap 75% of the letters in four-letter
words and 80% of the letters in five-letter words.
Our question is whether or not the absence of significant facilitation
for orthographically related pairs in the three contrasted studies (Drews
and Zwitserlood 1995; Frost et al. 1997; Rastle et al. 2000) and lack of
full priming (equal facilitation for orthographic pairs as for identity
priming) was the consequence of long targets differing in length from the
prime and with variable relation frequency. The inflectional morphologi-
cal priming manipulated in our two experiments permitted a better con-
straint on the orthographic matching and relative prime–target frequency
that might be responsible for the significant orthographic facilitation and
full priming at 32 ms. SOA, similar to the morphological facilitation. In
fact, both priming categories only begin to differ at 64 ms. SOA.

Semantic contrasts

The time-course processing for semantic activation is also different from

the orthographic and morphological activation: the semantic facilitation
occurred only at the last SOA (250 ms.), in contrast with the morphologi-
cal facilitation that is obtained from the earlier SOA (32 ms.). The
probability of semantic priming increases with the prime duration. De
Groot (1990) indicates that unmasked prime presentation is an effective
source of priming as a result of postlexical-meaning integration. However,
the data are not clear on this question. Some studies found facilitation
with very short masked priming (Durante and Hirshman 1994; Perea
and Gotor 1997). However, Frost et al. (1997) did not find semantic
priming at 43 ms. SOA using a masked procedure, whereas this target
was facilitated when preceded by a morphologically related word.
254 A. Dominguez, J. Segui, and F. Cuetos

This point is central for our proposal because one of our objectives
was to differentiate morphological processing from orthographic and
semantic processing. To reach this dissociation it is not sufficient to
obtain, as we obtained in the first experiment, a significant difference (at
64 ms. SOA) between morphological priming and orthographic or seman-
tic priming. The morphological facilitation (68 ms.) should be bigger
than the sum of orthographic plus semantic facilitation (28 ms.+
14 ms.=42 ms.). However, an ANOVA was carried out and the result
does not offer significant differences between the two sources of facilita-
tion at any of the three SOAs2 (see Balota and Paul 1996 for the applica-
tion of the specific procedure to a different semantic domain). Therefore,
the conclusion should be that morphological facilitation may be explained
without appealing to a specific treatment of morphological relatives, such
as lexical access by the root or morphological organization of the lexicon.
The semantic and orthographic overlap in morphological pairs is suffi-
cient to explain the extra-activation (facilitation) at 64 ms. SOA.
However, in this respect the study of Rastle et at. (2000) is very relevant.
The use of derivational complex words allows, in English, the introduc-
tion of a category of morphologically related words that are opaque with
respect to meaning (apartment–part). Rastle et al. found a similar facilita-
tion for morphologically related words in both morphological conditions:
semantically transparent (departure–depart) and opaque (apartment–
apart) at shorter SOA of 43 ms. This priming effect and the significant
difference from the facilitation produced by semantically plus orthograph-
ically nonmorphologically related words (screech–scream) in the second
experiment permitted them to affirm that morphological relations are not
merely a sum of orthographic and semantic activation.
Inflectional morphology does not allow semantic opacity between rela-
tives, and therefore we have not introduced a morphologically related,
nonsemantically related category. As a consequence, it is not possible to
know to what extent inflected gender words are treated in a different
manner from orthographically and semantically related words. However,
the general picture of our results is similar to that of Rastle et al., and
therefore converging evidence is expected in future research.
So, do our results allow the rejection of morphological processing of
the input? The answer is no. Our data do not provide sufficient evidence
to respond to this question. We may only affirm that there is a difference
in the time-course of the morphological pairs with respect to the semantic
and orthographic time-course. Some other studies have referred this
difference to morphological processing or to a central lexical organization
of morphological relatives (Drews and Zwitserlood 1995; Forster et al.
1987; Grainger et al. 1991; Rastle et al. 2000).
 Inflexional morphological priming 255

Nevertheless, two reservations should be made concerning our semantic

priming. To establish a correct base line of morphological priming, the
nature of the semantic relations should be similar in the morphological
pairs. This was not the case because the semantic pairs were synonymous
or associated words such as LUTO–pena ‘mourning’–‘sadness’, whereas
the morphological pairs were the gender of a person or animal as in
LOCA–loco ‘madwoman’–‘madman’. We carried out a post hoc analysis
to try to solve this problem.3 The associative strength of the related pairs
was introduced as a covariate, first, to neutralize the possible differences
in the semantic relation between morphological and semantic pairs, and
second, to compare the orthographic facilitation with the morphological
facilitation at 64 ms. SOA without a possible semantic influence.
Significant facilitation for the morphological pairs (68 ms. vs 2 ms.) was
again obtained under this procedure, which provides some evidence that
semantic pairs may be used in our experiments as a correct base line.

Conclusions

This research into Spanish inflexional morphology confirms the basic

pattern of data obtained in other languages. The different SOAs employed
across two experiments determined that morphology could operate at
64 ms. SOA, whereas orthography is computed earlier (35 ms.) and mean-
ing is later processed (250 ms.). The facilitation obtained for morphologi-
cally related words in all experiments and the time-course of this
activation could be interpreted as a specific way of access or lexical
organization for relatives. However, from a strict point of view this
facilitation does not surpass the sum of orthographic plus semantic
activation. Experiment 2 reveals that orthographic priming could be
equated to a repetition priming at 32 ms. SOA, whereas morphological
priming can occur at 645 ms. SOA.
The contrast with some other studies, in particular Rastle et al.’s study,
permits consideration of the following points:
1. The amount of orthographic (or formal ) priming may be underesti-
mated in other studies (Drews and Zwitserlood 1995; Frost et al. 1997;
Rastle et al. 2000). This point does not mean that morphological priming
is irrelevant for the lexical processor (in fact the amount of orthographic
overlapping in the morphological and orthographic categories of these
studies is well controlled).
2. The manipulation and control of semantic variables in Rastle et al.’s
study was crucial to separate morphological from semantic priming.
256 A. Dominguez, J. Segui, and F. Cuetos

3. The important linguistic differences between inflectional and deriva-

tional morphology described in the introduction of this article do not
promote meaningful differences between our results and Rastle et al.’s
study. Perhaps other functional properties of the stimuli, such as regular-
ity and irregularity, orthographic and semantic transparence, root fre-
quency and token frequency, etc., are responsible for the differences
found in other studies (Fowler et al. 1985).
4. Spanish is a language with a very productive inflectional system.
Suffixes of gender and number for nouns and adjectives and a high degree
of inflectional articulation in verbs (with almost a hundred different
flexion forms in each verbal root) outline the importance of inflectional
morphology in this language. English does not represent gender by a
suffix and verbs have reduced inflexion (Rastle et al. 2000). These cross-
linguistic differences, however, do not seem to be reflected in the results
of these experimental studies.

Received 14 June 2000 Universtiy of La Laguna, Tenerife

Revised version received C.N.R.S. and René Descartes University
4 December 2000 University of Oviedo, Asturias

Notes

* This research was supported by University of La Laguna 24162-97 and Dirección

General de Universidades grants to the first author. The second grant funded his
research for three months in the Laboratory of Experimental Psychology of the René
Descartes University at Paris, where this study began to be planned. Correspondence
should be addressed to Alberto Domı́nguez, Departamento de Psicologı́a Cognitiva,
Facultad de Psicologı́a, 38200 Tenerife, Spain. E-mail: adomin@ull.es.
1. The participants provided semantic associates to the prime words. Fifty low-frequency
four-letter words ending -a and -o (e.g. LUTO) were selected and presented to a group
of twenty subjects. They were asked to write down a semantically similar word beside
each word. From the pool of responses sixteen four-letter words were selected (e.g.
PENA), none of them sharing a letter in the same position as the prime, and with the
higher frequency of mention across subjects in the norms and printed frequency in the
Diccionario de Frecuencias de las Unidades Lingüı́sticas del Castellano [Frequency
Dictionary of Linguistic Units in Spanish] (Alameda and Cuetos 1995).
2. The analysis shows no significant differences in the 32 ms. SOA condition, F1(1, 39)
=0.06, MSe=8108, F2(1, 15)=0.37, Mse=6153.5; the 64 ms. SOA condition,
F1(1, 39)=1.59, MSe=7875,04, F2(1, 15)=0.96, Mse=3070.81; or the 250 ms. SOA
condition, F1(1, 39)=1.02, MSe=7740, F2(1, 15)=0.77, Mse=3360.5.
3. The primes of experiment 1 were given to 42 undergraduate students, who produced the
first word that they thought of. With this procedure we obtained an association index
for each of the prime–target pairs, and then an ANOVA, introducing the associative
strength for each item as a covariate, was carried out over the semantic, the morphologi-
cal, and the orthographic facilitation. No changes were obtained on the F∞ values for the
 Inflexional morphological priming 257

64 ms. SOA condition. Also, the significant difference between the morphological and
the orthographic facilitation was identical to that obtained without the introduction of
the associative strength as a covariate.

References

Alameda, J. R.; and Cuetos, F. (1995). Diccionario de Frecuencias de las Unidades

Lingüı́sticas del Castellano. Oviedo: Servicio de Publicaciones de la Universidad de
Oviedo.
Balota, D. A.; and Paul, S. T. (1996). Summation of activation: evidence from multiple
primes that converge and diverge within semantic memory. Journal of Experimental
Psychology: Learning, Memory and Cognition 22(4), 827–845.
Beauvillain, C.; and Segui, J. (1992). Representation and processing of morphological
information. In Orthography, Phonology, Morphology and Meaning, R. Frost and L. Katz
(eds.). Amsterdam. North-Holland.
Bentin, S.; and Feldman, L. B. (1990). The contribution of morphological and semantic
relatedness to repetition priming at short and long lags: evidence from Hebrew. Quarterly
Journal of Experimental Psychology 42A, 693–711.
Colé, P.; Beauvillain, C.; and Segui, J. (1989). On the representation and processing of
prefixed and suffixed derived words: a differential frequency effect. Journal of Memory and
Language 28, 1–13.
Colombo, L. (1986). Activation and inhibition with orthographically similar words. Journal
of Experimental Psychology: Human Perception and Performance 12, 226–234.
De Groot, A. M. B. (1990). The locus of the associative-priming effect in the mental lexicon.
In Comprehension Processes in Reading, D. A. Balota, G. B. Flores d’Arcais, and
K. Rayner (eds.). Hillsdale, NJ: Erlbaum.
Dominguez, A.; De Vega, M.; and Cuetos, F. (1997). Lexical inhibition from syllabic units
in Spanish visual word recognition Language and Cognitive Processes 12(4), 401–422.
Drews, E.; and Zwitserlood, P. (1995). Morphological and orthographic similarity in visual
word recognition. Journal of Experimental Psychology: Human Perception and Performance
21(5), 1098–1116.
Durante, R.; and Hirshman, E. (1994). Retrospective priming and masked semantic prim-
ing: the interfering effects of prime activation. Journal of Memory and Language 33,
112–127.
Feldman, L. B. (1994). Beyond orthography and phonology. Differences between inflections
and derivations. Journal of Memory and Language 33, 442–470.
—; and Andjelkovic, D. (1992). Morphological analysis in word recognition. In
Orthography, Phonology, Morphology and Meaning, R. Frost and L. Katz (eds.).
Amsterdam: North-Holland.
—; and Moskovljevic, J. (1987). Repetition priming is not purely episodic in origin. Journal
of Experimental Psychology: Learning, Memory and Cognition 15, 1–12.
Ferrand, L.; and Grainger, J. (1992). Phonology and orthography in visual word recogni-
tion: evidence from masked nonword priming. Quarterly Journal of Experimental
Psychology 45A, 353–372.
Forster, K. I. (1987). Form-priming with masked primes: the best match hypothesis. In
Attention and Performance XII, M. Coltheart (ed.). Hillsdale, NJ: Erlbaum.
—; and Davis, C. (1984). Repetition priming and frequency attenuation in lexical access.
Journal of Experimental Psychology: Learning, Memory, and Cognition 10, 680–698.
258 A. Dominguez, J. Segui, and F. Cuetos

—; and Davis, C.; Schoknecht, C.; and Carter, R. (1987). Masked priming with graphemi-
cally related forms: repetition or partial activation? Quarterly Journal of Experimental
Psychology 39A, 211–251.
Fowler, C. A.; Napps, S. E.; and Feldman, L. B. (1985). Relations among regular and
irregular morphologically related words in the lexicon as revealed by repetition priming.
Memory and Cognition 13, 241–251.
Frost, R.; Forster, K. I.; and Deutsch, A. (1997). What can we learn from the morphology
of Hebrew? A masked priming investigation of morphological representation. Journal of
Experimental Psychology: Learning, Memory and Cognition 23(4), 829–856.
Garcı́a-Albea, J. E.; Sanchez-Casas, R.; and Igoa, J. M. (1998). The contribution of word
form and meaning to language processing in Spanish: some evidence from monolingual
and bilingual studies. In Sentence Processing: A Cross-Linguistic Approach, D. Hilbert
(ed.). New York: Academic Press.
Grainger, J. (1990). Word frequency and neighborhood frequency effects in lexical decision
and naming. Journal of Memory and Language 29, 228–244.
—; Colé, P.; and Segui, J. (1991). Masked morphological priming in visual word recognition.
Journal of Memory and Language 30, 370–384.
Lukatela, G.; and Turvey, M. T. (1990). Automatic and pre-lexical computation of phonol-
ogy in visual word identification. European Journal of Cognitive Psychology 2(4), 325–343.
—; and Turvey, M. T. (1994a). Visual lexical access is initially phonological: 2. Evidence
of associative priming by words, homophones, and pseudohomophones. Journal of
Experimental Psychology: General 123, 107–128.
—; and Turvey, M. T. (1994b). Visual lexical access is initially phonological: 2. Evidence
from phonological priming by homophones and pseudohomophones. Journal of
Experimental Psychology: General 123(4), 331–353.
Marslen-Wilson, W. D.; Tyler, L. K.; Waksler, R.; and Older, L. (1994). Morphology and
meaning in the English mental lexicon. Psychological Review 101(1), 3–33.
McClelland, J. L.; and Rumelhart, D. E. (1981). An interactive activation model of context
effects in letter perception: part 1. An account of basic finding. Psychological Review
88, 375–407.
McQueen, J. M.; and Cutler, A. (1998). Morphology in word recognition. In The Handbook
of Morphology, A. Spencer and A. M. Zwicky (eds.). Oxford: Blackwell.
Napps, S. E. (1989). Morphemic relationships in the lexicon: are they distinct from semantic
and formal relationships? Memory and Cognition 17, 729–739.
Neely, J. H. (1991). Semantic priming effects in visual word recognition: a selective review
of current findings and theories. In Basic Processes in Reading: Visual Word Recognition,
D. Besner and G. Humphreys (eds.), 264–336. Hillsdale, NJ: Erlbaum.
Perea, M.; and Algarabel, S. (1992). Effects of neigborhood in masked repetition priming at
short SOAs. Paper presented at the Fifth Conference of the European Society for
Cognitive Psychology, Paris.
—; and Gotor, A. (1997). Associative and semantic priming effects occur at very short SOAs
in lexical decision and naming. Cognition 62, 223–240.
—; Gotor, A.; Rosa, E.; and Algarabel, S. (1995). Time course of semantic activation for
different prime–target relationships in the lexical decision task. Poster presented at the
36th Annual Meeting of the Psychonomic Society, November 10–12, Los Angeles.
Perfetti, C. A.; Bell, L. C.; and Delaney, S. M. (1988). Automatic (prelexical ) phonetic
activation in silent word reading: evidence from backward masking. Journal of Memory
and Language 27, 59–70.
—; and Tan, L. H. (1998). The time course of graphic, phonological and semantic activa-
tion in Chinese character identification Journal of Experimental Psychology: Learning,
Memory, and Cognition 24(1), 101–118.
 Inflexional morphological priming 259

Pinker, S. (1991). Rules of language. Science 253, 530–535.

Prasada, S.; and Pinker, S. (1993). Generalisation of regular and irregular morphological
processes. Language and Cognitive Processes 8, 1–56.
Rastle, K.; Davis, M. H.; Marslen-Wilson, W. D.; and Tyler, L. K. (2000). Morphological
and semantic effects in visual word recognition: a time-course study. Language and
Cognitive Processes 15(4/5), 507–537.
Sandra, D. (1994). The morphology of the mental lexicon: internal word structure viewed
from a psycholinguistic perspective. Language and Cognitive Processes 9(3), 227–269.
Segui, J.; and Grainger, J. (1990). Priming word recognition with orthographic neighbors:
effects of relative prime–target frequency. Journal of Experimental Psychology: Human
Perception and Performance 16, 65–76.
Seidenberg, M. S. (1987). Sublexical structures in visual word recognition: access units or
orthographic redundancy? In Attention and Performance XII: The Psychology of Reading,
M. Coltheart (ed.). Hillsdale, NJ: Erlbaum.
Stolz, J. A.; and Feldman, L. B. (1995). The role of orthographic and semantic transparency
of the base morpheme in morphological processing. In Morphological Aspects of Language
Processing, L. B. Feldman (ed.). Hillsdale, NJ: Erlbaum.
Taft, M. (1979). Recognition of affixed words and the word-frequency effect. Memory and
Cognition 7, 263–272.
—(1994). Interactive activation as a framework for understanding morphological process-
ing. Language and Cognitive Processes 9, 271–294.
—; and Forster, K. I. (1975). Lexical storage and retrieval of prefixed words. Journal of
Verbal Learning and Verbal Behavior 14, 638–647.

View publication stats