Content uploaded by Isidora Gatarić
Author content
All content in this area was uploaded by Isidora Gatarić on Jun 27, 2019
Content may be subject to copyright.
139
UDK: 81’23:[165.194:811.163.41’367.622
Originalni naučni rad
doi: 10.19090/pp.2019.2.139-156
DO MORPHOLOGICAL FEATURES AFFECT
THE COGNITIVE PROCESSING OF
DEVERBAL NOMINALS IN SERBIAN?2
The aim of this study was to examine whether different morpho-
logical characteristics of Serbian deverbal nominals affect their
lexical processing. According to morphological differences, there
are three subtypes of the process and result deverbal nominals in
Serbian: (i) result nominals end with the zero sufx, while process
nominals end with the deverbal sufx –nje (e.g., žubor/žuborenje
[eng. burble]); (ii) result nominals differ from process nominals in
the presence of the –va inx (e.g., rešenje/rešavanje [eng. so-
lution]); (iii) process nominals end with the deverbal sufx –nje,
while result nominals end with other derivational sufxes (e.g.,
rotiranje/rotacija [eng. rotation]). The nal results of three self–
paced reading experiments suggest that different morphological
features do not affect the processing of deverbal nominals, which
strongly supports a–morphous approach to the morpho–lexical
processing, as well as the distributed morphology perspective in
the eld of theoretical linguistics.
Key words: a–morphous morphology, derived nouns, deverbal
nominalization, distributed morphology, morpho–lexical process-
ing
th
th International Morphology Meeting
th International Morphological Processing
Isidora Gatarić1
Social Sciences and
Computing, University
of Belgrade
Sanja Srdanović
Anja Šarić
Department of
Linguistics, Goethe
University in Frankfurt
Corresponding author e–mail:
gataric.isidora@gmail.com
PRIMENJENA PSIHOLOGIJA, 2019, Vol.
12
(2), STR. 139-156
Primljeno: 30. 03. 2019.
Primljena korekcija:
06. 06. 2019.
Prihvaćeno za štampu:
13. 06. 2019.
primenjena psihologija 2019/2
Isidora Gatarić, Sanja Srdanović, and Anja Šarić
140
Introduction
-
-
examination in the example
The instructor’s examination of the student
-
verbal nominal exam in the example The instructor’s exam-
examination in the example The instructor’s exami-
nation
-
simple event nominal category is considered to behave identically as process de-
distinctive categories of deverbal nominals are relevant for the Serbian language:
drhtaj in the example
drhtaj je nagoveštavao dolazak zime [eng. -
al of winter
drhtanje in the example [eng.
of deverbal nominals is that the process deverbal nominals take obligatory argu-
on the richness of morphology of the language in question.
Morphological Distinction between Deverbal Nominals in Serbian
them perfect candidates for research of morphological effects in morpho–lexical
primenjena psihologija, str. 139-156
DO MORPHOLOGICAL FEATURES AFFECT THE COGNITIVE PROCESSING OF DEVERBAL NOMINALS IN SERBIAN?
141
processing. Previous studies interested in the theoretical explanation of deverbal
nominalization in Serbian suggest that this phenomenon is morphologically very
–nje–ba, –ija, –aj, –nja, –ak-
–va
three subtypes of process and result deverbal nominals can be distinguished in
-
–njelet/letenje [eng.
and process deverbal nominals differ in the presence of the –varešenje/
rešavanje [eng. solving–nje
rotiranje/rotacija
[eng. rotation
no previous empirical studies interested in the examination of deverbal nominals
processing with respect to these morphological differences.
Cognitive Processing of Deverbal Nominals
Having in mind the fact that this phenomenon has been intriguing to re-
a minimal interest has been devoted to deverbal nominalization in the domain of
of syntactic complexity of deverbal nominals on the processing of entire sentenc-
results of this study suggested that sentences with the deverbal nominals with
simpler linguistics structure were processed faster than those with the complex
results suggesting that the more complex the syntactic structure of deverbal nom-
and syntactic complexities played the dominant roles in the cognitive process-
the syntactic and semantic differences of process and result deverbal nominals
-
-
-
mantically and syntactically more complex deverbal nominals (process deverbal
nominals) were processed slower than those with the simpler linguistic structure
-
primenjena psihologija 2019/2
Isidora Gatarić, Sanja Srdanović, and Anja Šarić
142
with the more complex linguistic structure were rated as less acceptable accord-
conclude that participants perhaps would need more time to process deverbal
-
-
them control (and vary) only for the syntactic and semantic features of deverbal
-
cognitive processing of process and result deverbal nominals.
Different Perspectives in the Processing of Morphologically Complex
Words
The morphological complexity of words has intrigued and inspired re-
the cognitive processing of morphologically complex words in any language. On
-
ing to the authors of these models morphemes are represented as independent
-
of authors who propose an a–morphous perspective in the morpho–lexical pro-
cessing. They suggest that morphemes do not play an important independent
proposed a few models for the interpretation of results observed in the empiri-
-
-
primenjena psihologija, str. 139-156
DO MORPHOLOGICAL FEATURES AFFECT THE COGNITIVE PROCESSING OF DEVERBAL NOMINALS IN SERBIAN?
143
successfully explains various morpho–lexical phenomena in different languag-
-
-
logically complex words many years before previously mentioned a–morphous
-
-
tion that a–morphous perspective in the lexical processing is very similar to dis-
existence of the single morpheme characteristics effects in the lexical process-
The Present Study
Guided by the fact that there are no previous similar studies dealing with this
-
logical features of Serbian deverbal nominals affected their processing. The mor-
–nje
differ in the presence of the –va
–nje
the second aim of this study was to answer which of the two perspectives in the
-
Experiment 1
-
phological features of Serbian deverbal nominals affected their processing. In the
-
–nje
primenjena psihologija 2019/2
Isidora Gatarić, Sanja Srdanović, and Anja Šarić
144
were more complex in nature than the result deverbal nominals. The stimuli were
presented in the sentence context in the self–paced reading task in all three ex-
studies about the importance of sentences context in the examination of morpho–
Method
Participants. Participants in this experiment were undergraduate students
N
-
-
versity of Novi Sad)3
or corrected–to–normal vision.
Stimuli. The stimuli in this experiment were 48 sentences with the Serbian
-
the experiment had its pair: 24 result nominals ended with the zero morpheme
[eng. burble-
–nje [eng. burble]).4 The pairs of deverbal nominals were used
in order to control the effects that could arise from the characteristics of a stem
/ [eng.
burble
Jovana je [eng.
is[eng. heard
[eng. burble -
vode [e.g. water]) as il-
lustrated in (1).
1a vode.
1b vode.
[eng. Jovana heard the burble of water.]
identical sentences (for one pair of deverbal nominals) were selected as stimuli in
Committee of that institution.
primenjena psihologija, str. 139-156
DO MORPHOLOGICAL FEATURES AFFECT THE COGNITIVE PROCESSING OF DEVERBAL NOMINALS IN SERBIAN?
145
conducted research in Serbian.
Design. The factor that was manipulated in this experiment was the type
of deverbal nominals (process or result) that differed in some morphological
–nje
lemma frequency were included as the control variables. The lemma frequencies
were retrieved from srWac
length was calculated according to the number of letters. The dependent variable
in this experiment was the reading time of deverbal nominals (measured in mil-
liseconds).
Procedure. The stimuli were presented in a self–paced reading task created
in the software OpenSesame
-
-
ticipants were verbally instructed to read the words presented at the screen as
ENTER (on the keyboard) when they read
-
tions (about the previous sentence) were given on the screen on several occa-
-
presentation was randomized for each participant.
Results
-
in free statistical software R mgcv
itsadug-
word length were standardized by centring to zero and dividing by the stand-
-
-
primenjena psihologija 2019/2
Isidora Gatarić, Sanja Srdanović, and Anja Šarić
146
-
effect of stimuli and the random effect of participants. The random effect of par-
ticipants was included with by–participant factorial smooths over trials from the
result from the trials characteristics in the case of different participants from the
-
model criticism was applied to the model following the procedure proposed by
Table 1.
Table 1
-
sponse latencies from Experiment 1
t Pr(>|t|)
Intercept 6.23 .04 .00***
Type of deverbal nominals = result –.04 .02 –1.62 .10
Trial order –.00 .00 –.63
.01 .01 1.31 .18
–.01 .00 .02
Smooth terms edf Ref.df F p
s(Stimuli) 4.18 44 .10
s 7.33 .00***
Notes. s – thin plate regression spline smooth.
*** p < .0001.
which means that longer deverbal nominals are processed slower than shorter
-
certain morphological differences between Serbian process and result nominals
primenjena psihologija, str. 139-156
DO MORPHOLOGICAL FEATURES AFFECT THE COGNITIVE PROCESSING OF DEVERBAL NOMINALS IN SERBIAN?
147
Experiment 2
-
–va
Method
Participants.
speakers with normal or corrected–to–normal vision. None of the participants
Stimuli.
-
–vanje) were collected: result
deverbal nominals (N –va[eng. cut
while the process deverbal nominals (N–va
[eng. cut]). The stimuli sentences were created with the identical syntactic regula-
1a struje.
1b struje.
[eng. The Electric Distribution Company announced power cuts.]
Design and Procedure. The two–level factor was a type of deverbal nomi-
–va. The same control variables and dependent
Results
-
primenjena psihologija 2019/2
Isidora Gatarić, Sanja Srdanović, and Anja Šarić
148
the whole process of the preparation of the data for the statistical analysis was the
-
Table 2.
Table 2
-
sponse latencies from the Experiment 2
t Pr(>|t|)
Intercept 124.10 .00***
Trial order (order of presentation) –.00 .00 –.61
Type of deverbal nominals = result –.04 .02 .07
.01 .00 1.37 .16
Smooth terms edf Ref.df F p
s 1.00 .13
s 2.43 .12
s(Stimuli) .00 .61
s .00***
Notes.s – thin plate regression spline
smooth.
*** p < .0001.
explanations for the lack of this effect can be that the stimuli in this experiment
-
-
primenjena psihologija, str. 139-156
DO MORPHOLOGICAL FEATURES AFFECT THE COGNITIVE PROCESSING OF DEVERBAL NOMINALS IN SERBIAN?
149
Experiment 3
-
–nje
–ba, –ija, –aj, –nja, –ak, –idba).
Method
Participants. Sixty–eight undergraduate students (mostly female) from the
-
normal vision. None of the participants participated in the previous two experi-
ments.
Stimuli. The sentences (N = 48) with the pairs of process and result deverbal
–aj in deverbal noun premeštaj [eng. relocation-
–njepremeštanje [eng. reloca-
tion
created with the identical syntactic regulations
1a premeštaj kancelarije.
1b premeštanje kancelarije.
[eng. ]
Design and Procedure.
Results
-
-
tical analysis was the same as in the previous experiments. The same statistical
model is presented in the Table 3.
primenjena psihologija 2019/2
Isidora Gatarić, Sanja Srdanović, and Anja Šarić
150
Table 3
-
sponse latencies from the Experiment 3
t Pr(>|t|)
Intercept 6.26 .00***
Trial order (order of presentation) .00 .00 .17
Type of deverbal nominals = result .00 .01 .18
Smooth terms edf Ref.df F p
2.88 6.27 .00**
1.14 1.22 .68
s(Stimuli) 44 .01*
7.07 .00***
Notes. s – thin plate regression spline smooth.
* p < .01. ** p < .001. *** p < .0001.
which means that longer deverbal nominals are being processed slower. The ef-
as well as the main effect of the type of deverbal nominals. This suggests that
certain morphological features that vary in this experiment do not contribute to
the appearance of a difference in the processing of process and result deverbal
nominals.
Discussion
The current research was primarily aimed at examining whether different
morphological characteristics of deverbal nominals affected their lexical process-
ing in Serbian. Three experiments with the self–paced reading tasks were carried
out in order to get an answer to this research question. Morphological differences
-
periment 1 were sentences with the following type of deverbal nominals: result
–nje[eng. burble
stimuli were result and process deverbal nominals that differed in the presence
of the –va rešenje/rešavanje [eng. solution
primenjena psihologija, str. 139-156
DO MORPHOLOGICAL FEATURES AFFECT THE COGNITIVE PROCESSING OF DEVERBAL NOMINALS IN SERBIAN?
151
–nje-
rotiranje/rotacija [eng. rotation]). Another aim of this research was to
answer which of the two perspectives in the morpho–lexical processing would be
the most appropriate for the explanation of this phenomenon. The data analysis
of all three experiments suggests that there is no effect of morphological charac-
teristics in the cognitive processing of deverbal nominals in the Serbian language.
well as the distributed morphology perspective from the theoretical linguistics.
Semantic and syntactic complexity of deverbal nominals drew attention of a
number of language scientists who dealt with both theoretical and empirical re-
-
empirical studies were interested only in the syntactic and semantic effects on
cognitive processing of deverbal nominals. Almost all of these studies suggested
identical results that the syntactic and semantic complexity of deverbal nominals
-
acteristics of deverbal nominals on their processing. Although theoretical studies
in Serbian show that there was a certain morphological complexity of deverbal
-
ogy itself was not the subject of empirical research interested in the processing
a discussion that has been going on for years in the morphological research cir-
cle is whether morphemes themselves affect processing time of the whole word
-
-
-
ences in the processing of two types of deverbal nominals that differ because the
–nje
-
primenjena psihologija 2019/2
Isidora Gatarić, Sanja Srdanović, and Anja Šarić
152
single morphemes characteristics in the processing of deverbal nominals. One of
that there are no differences in the processing time of derived nouns with zero–
–nje). This could
be one of the most prominent pieces of evidence that a–morphous perspective
processing of deverbal nominals (the one that is highly controlled in this study).
-
ferences in the processing of process and result nominals that differ in the pres-
–va. Those results also support the a–morphous perspective in
-
-
-
-
Bybee
Conclusion
processing of deverbal nominals in Serbian is not affected by the morphological
previous studies that highlight the importance of semantic and syntactic differ-
the morphological features of deverbal nominals are not of crucial importance for
the appearance of differences in the processing time of process and result dever-
therefore provoking the traditional view in the morpho–lexical processing.
primenjena psihologija, str. 139-156
DO MORPHOLOGICAL FEATURES AFFECT THE COGNITIVE PROCESSING OF DEVERBAL NOMINALS IN SERBIAN?
153
Acknowledgements
th
th International Morphology Meeting (Buda-
th
-
per.
References
(Part I: The Nominalization Puzzle). Language and Linguistics Compass, 4
A–morphous morphology. Cambridge: Cambridge University
Press.
Analyzing Linguistic Data: A Practical Introduction to Statis-
tics Using R. United States: Cambridge University Press.
Analyzing Reaction Times. International Journal
of Psychological Research, 3
Bra-
zilian Journal of Applied Linguistics, 11,
Regression Diagnostics. Identifying
-
ability and Mathematical Statistics.
The
Mental Lexicon, 6
Morphology: A Study of the Relation Between Meaning and
Form. Amsterdam: John Benjamins Publishing Company.
The cognitive pro-
cessing od process and result deverbal nominals in English and Serbian. Paper
th
-
Primenjena psihologija, 12
Data Analysis Using Regression and Multilevel/Hi-
erarchical Models. Cambridge: Cambridge University Press. doi:10.1017/
Argument Structure. Cambridge: The MIT Press.
Proceedings of NELS 20 (pp.
primenjena psihologija 2019/2
Isidora Gatarić, Sanja Srdanović, and Anja Šarić
154
-
MITWPL 30: Papers at the Interface (pp.
Proceedings of Sinn und Bedeutung, 17,
Kognitivna psihologija [Cognitive psychology].
-
Linguistic Inquiry, 30,
The Serbian web corpus srWaC
graphical experiment builder for the social sciences. Behavior Research Meth-
ods,
Slavic and Romance languages. Linguistik Online, 77
On the processing of thematic features in deverbal nomi-
nals
Memory and Cog-
nition, 5
-
tion in lexical decision. PLoS ONE, 12doi:10.1371/journal.
-
sible data science in language research. Cognitive Linguistics, 27
Cross–linguistic Investigations of Nominalization Patterns. Amster-
dam: John Benjamins Publishing Company.
Word knowledge and word usage: A cross–disciplinary guide to the
mental lexicon
R Core Team (2017). R: A language and environment for statistical computing.
http://
www.R–project.org/
Cognitive processing of result and process nominals in English
and Serbian -
versity of Novi Sad.
primenjena psihologija, str. 139-156
DO MORPHOLOGICAL FEATURES AFFECT THE COGNITIVE PROCESSING OF DEVERBAL NOMINALS IN SERBIAN?
155
of Research. Psychological Bulletin, 124 doi:10.1037//0033–
CogSci,
Theoretical and empirical vi-
dimus of the simple event nominals (non) existence in Serbian. Paper presented
th
-
cy effect. Quarterly Journal of Experimental Psychology, 57A
Journal of Verbal Learning and Verbal Behavior, 14, 638–647. doi:10.1016/
itsadug: Interpreting Time
Series and Autocorrelated Data Using GAMMs. R package version 1.0.1.
Generalized Additive Models: An Introduction with R. United
States: Chapman and Hall/CRC.
-
hood estimation of semiparametric generalized linear models. Journal of the
Royal Statistical Society, 73
The structure of the Serbian noun phrase (Unpublished doctoral
dissertation). United States: University of Texas at Austin.
primenjena psihologija 2019/2
Isidora Gatarić, Sanja Srdanović, and Anja Šarić
156
DA LI MORFOLOŠKE ODLIKE UTIČU NA
KOGNITIVNU OBRADU DEVERBALNIH
IMENICA U SRPSKOM JEZIKU?
Cilj ove studije bio je da se ispita da li različite morfološke ka-
rakteristike srpskih deverbalnih imenica utiču na njihovu leksi-
čku obradu. Prema morfološkim razlikama, postoje tri podvrste
procesnih i rezultativnih deverbalnih imenica u srpskom jeziku:
(i) rezultativne imenice koje se završavaju nultim suksom, dok
se procesne imenice završavaju deverbalnim suksom –nje (npr.
žubor/žuborenje [eng. burble]); (ii) rezultativne imenice se razli-
kuju od procesnih u prisustvu inksa –va (npr. rešenje/rešavanje
[eng. solution]); (iii) procesne imenice se završavaju deverbalnim
suksom –nje, dok se rezultativne završavaju nekim drugim deri-
vacionim suksima (npr. rotiranje/rotacija [eng. rotation]). Finalni
rezultati tri eksperimenta sa zadatkom čitanja slobodnim tempom
pokazuju da različite morfološke odlike ne utiču na obradu de-
verbalnih imenica, što podržava a–morfni pristup morfo–leksičkoj
obradi, kao i distributivno–morfološku perspektivu iz oblasti teo-
rijske lingvistike.
Ključne reči: a–morfna morfologija, derivirane imenice, dever-
balna nominalizacija, distributivna morfologija, morfo–leksička
obrada
Isidora Gatarić
Računarstvo u
društvenim naukama,
Univerzitet u
Beogradu
Sanja Srdanović
Anja Šarić
Odsek za lingvistiku,
Goethe Univerzitet u
Frankfurtu