ChapterPDF Available

Expressing and categorizing motion in French and English: Verbal and non-verbal cognition across languages


Abstract and Figures

Language-specific properties influence motion expression (Slobin 2004; Talmy 2000), but it is still debated whether they also influence non-verbal spatial cognition. We compare how English and French speakers perform three tasks involving motion events: non-verbal categorization based on cartoons during a dual task (articulatory suppression); verbal categorization involving target sentences; and descriptions of motion events. Descriptions show more manner expression in English and variation in both groups as a function of event properties. However, both groups frequently choose Path as categorization criterion, particularly in verbal categorization, but language effects do show that English speakers are more sensitive to event properties. Thus, typology has a strong impact on verbal cognition but also a partial influence on non-verbal cognition that must be taken into account in future research on the relation between language and thought
Content may be subject to copyright.
2nd proofs
Motion and Space across Languages
2nd proofs
Human Cognitive Processing (HCP)
Cognitive Foundations of Language Structure and Use
ISSN 1387-6724
Volume 59
Motion and Space across Languages. eory and applications
Edited by Iraide Ibarretxe-Antuñano
Editorial Board
Bogusław Bierwiaczonek
Jan Dlugosz University, Czestochowa, Poland /
Higher School of Labour Safety Management,
Mario Brdar
University of Osijek, Croatia
Barbara Dancygier
University of British Columbia
N.J. Enfield
University of Sydney
Elisabeth Engberg-Pedersen
University of Copenhagen
Ad Foolen
Radboud University Nijmegen
Raymond W. Gibbs, Jr.
University of California at Santa Cruz
Rachel Giora
Tel Aviv University
Elżbieta Górska
University of Warsaw
Martin Hilpert
University of Neuchâtel
Zoltán Kövecses
Eötvös Loránd University, Hungary
Teenie Matlock
University of California at Merced
Carita Paradis
Lund University
Günter Radden
University of Hamburg
Francisco José Ruiz de Mendoza Ibáñez
University of La Rioja
Doris Schönefeld
University of Leipzig
Debra Ziegeler
University of Paris III
is book series is a forum for interdisciplinary research on the grammatical
structure, semantic organization, and communicative function of language(s), and
their anchoring in human cognitive faculties.
For an overview of all books published in this series, please see
Klaus-Uwe Panther
University of Hamburg
Linda L. ornburg
2nd proofs
Motion and Space
across Languages
eory and applications
Edited by
Iraide Ibarretxe-Antuñano
University of Zaragoza
John Benjamins Publishing Company
Amsterdam / Philadelphia
2nd proofs
Table of contents
Editor and contributors xi
Past, present, and future of motion research 1
Leonard Talmy
Motion and semantic typology: A hot old topic with exciting caveats 13
Iraide Ibarretxe-Antuñano
PartI. Delving into motion event typology
e typology of manner expressions: A preliminary look 39
Kimi Akita
Expressing and categorizing motion in French and English:
Verbal and non-verbal cognition across languages 61
Maya Hickmann, Helen Engemann, Efstathia Soroli,
Henriëtte Hendriks and Coralie Vincent
e functional nature of deictic verbs and the coding patterns of Deixis:
An experimental study in English, Japanese, and ai 95
Yo Matsumoto, Kimi Akita and Kiyoko Takahashi
e importance of minority languages in motion event typology:
e case of Aragonese and Catalan 123
Iraide Ibarretxe-Antuñano, Alberto Hijazo-Gascón
and María-Teresa Moret-Oliver
Latin to Ancient Italian motion constructions: A complex typological shi 151
Monica Mosca
2nd proofs
viii Motion and Space across Languages
e early life of borrowed path verbs in English 177
Judith Huber
Non-actual motion in language and experience 205
Johan Blomberg
Metaphorical motion constructions across specialized genres 229
Rosario Caballero
PartII. Expanding motion event typology
Crossing the road or crossing the mind: How dierently do we move
across physical and metaphorical spaces in speech and in gesture? 257
Şeyda Özçalışkan, Lauren J. Stites and Samantha N. Emerson
inking for speaking about motion in a second language:
Looking back and forward 279
Teresa Cadierno
Motion events contrasts in Romance languages: Deixis in Spanish
as a second language 301
Alberto Hijazo-Gascón
Verb framed, satellite framed or in between? A L2 learner’s thinking
for speaking in her L1 and L2 over 14 years 329
Gale Stam
On the reception of translations: Exploring the impact
of typological dierences on legal contexts 367
Ana Rojo and Paula Cifuentes-Férez
Applying language typology: Practical applications of research
on typological contrasts between languages 399
Luna Filipović
2nd proofs
Table of contents ix
Typologies and language use 419
Dan I. Slobin
Name index 447
Subject index 4??
Language index 4??
2nd proofs
Editor and contributors
Iraide Ibarretxe-Antuñano
Universidad de Zaragoza
Facultad de Filosofía y Letras
Departamento de Lingüística General e Hispánica
Pedro Cerbuna, 12
E-50009 Zaragoza
Kimi Akita
Nagoya University
Graduate School of Humanities
Furo-cho, Chikusa-ku, Nagoya-shi,
Aichi 464-8601
Johan Blomberg
Centre for Languages and Literature,
Lund University
Box 201
221 00 Lund
Rosario Caballero
Universidad de Castilla-La Mancha
Facultad de Letras
Avda. Camilo José Cela, s/n
13071 Ciudad Real
Teresa Cadierno
University of Southern Denmark
Department of Language and
Campusvej 55
5230 Odense M
2nd proofs
xii Motion and Space across Languages
Paula Cifuentes
Universidad de Murcia
Departamento de Traducción e
Facultad de Letras
Campus Universitario de La Merced
Santo Cristo, 1
30001 Murcia
Samantha N. Emerson
Georgia State University
Department of Psychology
PO Box 5010
Atlanta, GA 30302-5010
Helen Engemann
Universität Mannheim
Anglistische Linguistik
Schloss EW 266D-68131 Mannheim
Luna Filipović
University of East Anglia
Faculty of Arts and Humanities
School of Politics, Philosophy, Language
and Communication Studies
Norwich Research Park
Norwich, NR4 7TJ
United Kingdom
Henriëtte Hendriks
University of Cambridge
Faculty of Modern and Medieval
Dept. of eoretical and Applied
9, West Road
Cambridge CB3 9DP
United Kingdom
Maya Hickmann
Centre CNRS Pouchet
Laboratoire Structures Formelles du
Langage (UMR 7023)
59, rue Pouchet
F-75017 Paris
Alberto Hijazo-Gascón
University of East Anglia
Faculty of Arts and Humanities
School of Politics, Philosophy, Language
and Communication Studies
Norwich Research Park
Norwich, NR4 7TJ
United Kingdom
Judith Huber
Erlangen-Nürnberg (FAU)
Institut für Anglistik und
Bismarckstr. 191054 Erlangen
2nd proofs
Editor and contributors xiii
Iraide Ibarretxe-Antuñano
Universidad de Zaragoza
Facultad de Filosofía y Letras
Departamento de Lingüística General
e Hispánica
Pedro Cerbuna, 12
E-50009 Zaragoza
Yo Matsumoto
Kobe University
Department of Linguistics
1-1 Rokkodai-cho, Nada-ku
Kobe 657-8501
María Teresa Moret-Oliver
Universidad de Zaragoza
Facultad de Filosofía y Letras
Departamento de Lingüística General e
Pedro Cerbuna, 12
E-50009 Zaragoza
Monica Mosca
Università del Piemonte Orientale
Amedeo Avogrado”
Dipartimento di Studi Umanistici
Via G. Ferraris, 116
13100 Vercelli VC
Şeyda Özçalışkan
Georgia State University
Department of Psychology
PO Box 5010
Atlanta, GA 30302-5010
Ana Rojo
Universidad de Murcia
Departamento de Traducción e
Facultad de Letras
Campus Universitario de La Merced
Santo Cristo, 1
30001 Murcia
Dan I. Slobin
Professor Emeritus of Psychology and
University of California, Berkeley
Efstathia Soroli
Laboratoire “Savoirs, Textes, Langage
(UMR 8163)
University of Lille 3
Rue du Barreau – BP 60149, 59653
Villeneuve dAscq Cedex
Lauren J. Stites
Georgia State University
Department of Psychology
PO Box 5010
Atlanta, GA 30302-5010
Gale Stam
National Louis University
Department of Psychology
5202 Old Orchard Road, Suite 300
Skokie, IL 60077
2nd proofs
xiv Motion and Space across Languages
Kiyoko Takahashi
Kanda University of International Studies
Faculty of Foreign Languages
Department of Asian Languages
1-4-1 Wakaba, Mihama-ku, Chiba-shi,
Chiba 261-0014
Leonard Talmy
Professor Emeritus of Linguistics and
Director Emeritus of the Center for
Cognitive Science
University at Bualo,
State University of New York
Coralie Vincent
Centre CNRS Pouchet
Laboratoire Structures Formelles du
Langage (UMR 7023)
59, rue Pouchet
F-75017 Paris
2nd proofs
Expressing and categorizing motion
in French and English
Verbal and non-verbal cognition across languages
Maya Hickmanni, Helen Engemannii, Efstathia Soroliiii,
Henriëtte Hendriksiv and Coralie Vincenti
iCNRS & University of Paris 8 / iiMannheim University /
iiiUniversity of Lille 3 / ivUniversity of Cambridge
Language-specic properties inuence motion expression (Slobin 2004; Talmy
2000), but it is still debated whether they also inuence non-verbal spatial
cognition. We compare how English and French speakers perform three tasks
involving motion events: non-verbal categorization based on cartoons during
a dual task (articulatory suppression); verbal categorization involving target
sentences; and descriptions of motion events. Descriptions show more manner
expression in English and variation in both groups as a function of event prop-
erties. However, both groups frequently choose Path as categorization criterion,
particularly in verbal categorization, but language eects do show that English
speakers are more sensitive to event properties. us, typology has a strong
impact on verbal cognition but also a partial inuence on non-verbal cognition
that must be taken into account in future research on the relation between lan-
guage and thought.
Keywords: categorization, lexicalization, Manner (of motion), Path (of motion),
spatial cognition, typology
1. Introduction
Spatial systems show considerable crosslinguistic variation which has been studied
within a typological perspective (Slobin 2004; Talmy 2000). Although crosslin-
guistic variation has been reported in studies focusing on language production,
a most debated question remains, namely whether language-specic factors also
have a deeper impact on speakers’ non-verbal cognition (e.g. Blomberg and Zlatev
2009; Filipović 2011; Gennari et al. 2002; Majid et al. 2004; Papafragou and Selimis
 ./hcp..hic
©  John Benjamins Publishing Company
2nd proofs
62 Maya Hickmann et al.
2010; Soroli 2011, 2012). is question constitutes perhaps one of the most dicult
challenges revolving around a number of controversies that are far from being
resolved. Furthermore, addressing this question requires complex experimental
designs and complementary methodologies. e present study aims to contribute
some answers to these debates by comparing verbal and non-verbal responses
produced by adult native speakers in two language groups, English and French,
diering with respect to their lexicalization patterns for motion expression.
1.1 Motion expression across languages
Tal my (200 0) ha s propos ed a t ypolog y t hat dist ing ui shes sever al groups of l an-
guages according to their lexicalization patterns (also see Ibarretxe-Antuñano this
volume). Among them, as shown in (1)–(4), “satellite-framed languages” (hereaer
“S-languages”, e.g. English) lexicalize Manner in the verb root and express Path in
other elements outside of the verb, while “verb-framed languages” (“V-languages”,
e.g. French) lexicalize Path in the verb root and express Manner by peripheral
means, if at all.
(1) to walk, run, jump across, up, down, into, away
(2) traverser ‘to cross’, monter ‘to ascend’, descendre ‘to descend’, entrer ‘to enter’,
partir ‘to leave en marchant ‘by walking’, en courant ‘by running’, en sautant
‘by jumping’
(3) She walked up
(4) Elle est montée
‘She ascended’
Several points should be noted with respect to this dichotomy. First, lexicalization
patterns dierentiate language groups when motion implies a change of location
as in (1)–(4) above. However, when motion takes place within a general location,
as in (5)–(6), S- and V-languages do not dier from one another (Manner in verb).
Furthermore, changes of location include dierent types of events that dier with
respect to spatio-temporal properties and may lead to some variation in motion
expression. us, motion in (7) intrinsically implies a boundary-crossing, i.e. a
“categorical” change from one location (outside) to another (inside). In contrast,
(8) implies a “gradual” change from a starting point to successively higher points
with no intrinsic endpoint (notwithstanding inferences, e.g. the top of the tree).
1. We prefer to avoid terms such as “satellites” or “adjuncts” to allow for a wider understanding
of all elements that may contribute to the expression of motion in the periphery of the verb root.
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 63
Similarly, (9) implies a categorical change with two potential boundaries (the two
sides of the street), whereas (10) does not.
(5) to walk, run in, on
(6) marcher, courir dans, sur
(7) e man ran into the house
(8) e squirrel ran up the tree
(9) e boy ran across the street
(10) e boy ran along the street
Second, these dierent lexicalization patterns are predominant and most typical
for a given language but do not exclude other means of expressing motion. us,
a V-like pattern is possible in English (path verb in (11)) when Manner is irrele-
vant in discourse. Conversely, a S-like pattern is possible in French (manner verb
in (12)) when Manner is marked and highlighted in discourse. In this respect,
boundary-crossings (e.g. , …) have a special status within the typology
since they are most likely to require a path verb in V-languages as compared to oth-
er types of events (e.g. , , …) (Aske 1989; Özçalıkan 2009, 2015; Slobin
2004, 2006). us, Example(13) is most likely to be interpreted as meaning that
the event (running) took place within a general location (the garage). Although a
reading implying a boundary-crossing (from outside to inside the garage) is not
impossible, it is highly unlikely, structure (14) being necessary and most natural
if such a reading is to be unambiguously expressed.
(11) Don’t cross the street!
(12) Il était en retard et il a couru jusqu’à l’école
‘He was late and he ran all the way to school
(13) Il a couru dans le garage
‘He ran in/into the garage
(14) Il est entré dans le garage en courant
‘He entered in/into the garage by running’
ird, some languages (e.g. Chinese, ai) do not quite t into the proposed dis-
tinction, suggesting that the typology may be rephrased in terms of a continuum
rather than a dichotomy between S- and V-languages (Slobin 2004). Yet another
question concerns the precise nature of the dierence between S- and V-languages.
According to Talmy, Path is the most basic component dening motion, the ex-
pression of which constitutes the major source of typological variation across lan-
guages (Path-in-verb vs. Path-in-satellites). Similarly, using a continuum rather
2nd proofs
64 Maya Hickmann et al.
than a dichotomy, Ibarretxe-Antuñano (2009) proposes to place languages on
a “Path-salience cline” depending on the degree to which they highlight path
information. By comparison, Slobin (2006) considers manner expression as rep-
resenting the major dierence in the typology (manner-in-verb for S-languages,
peripheral and/or optional for V-languages, also see Slobin this volume). Slobin
et al. (2014) further show that typological properties result in manner expressions
with variable lexical diversity. Analyses of particular manner properties in extend-
ed motion descriptions, such as ‘walking’, ‘running’, ‘non-canonical gaits’ (divided
into ‘bounce-and-recoil’ and ‘syncopated movements’), and ‘quadrupedal move-
ments’ (crawling), show that S-framed languages display much more granularity
in manner expression than V-languages, allowing more varied and ne-grained
manner distinctions (also see Soroli 2011 for similar results).
1.2 Verbal and non-verbal spatial cognition
On the basis of systematic crosslinguistic dierences, the last decades have wit-
nessed a revival of the Whorfian view of language use (Evans and Levinson
2009; Gumperz and Levinson 1996; Hickmann and Robert 2006; Levinson 2003;
Levinson and Wilkins 2006; Lucy 1992). New versions of this view (particularly
Slobin 1996, 2004) have proposed that our language (or language type) acts as a
lter” inviting us to focus our attention more on some aspects of reality than on
others when speaking. Much empirical evidence has supported this idea, show-
ing that language-specic properties partially determine how speakers select and
organize motion information when communicating, e.g., they pay more attention
to manner information in S-languages than in V-languages. Such language eects
have been observed in production studies across many languages and types of
speakers, including native adult speakers, monolingual children (e.g. Allen et al.
2007; Bowerman and Choi 2003; Choi and Bowerman 1991; Choi 2011; Harr 2012;
Hickmann et al. 2009b; Hickmann 2010), bilingual children acquiring two lan-
guages simultaneously (Engemann 2012), and adults acquiring a foreign language
(Cadierno 2004, 2008; Cadierno and Lund 2004; Carroll et al. 2012; Hendriks and
Hickmann 2011; Hendriks et al. 2008; von Stutterheim and Nüse 2003).
A more controversial question concerns the extent to which language-specic
factors may have deeper eects on human cognition, inuencing our event repre-
sentations not only when we are speaking but also in non-verbal tasks. Language
eects in verbal tasks are not sucient to support the claim that linguistic fac-
tors inuence cognition. Such eects can be seen as supercial, only showing
that language properties aect language use. Stronger claims of linguistic deter-
minism require language eects in behaviors that do not imply explicit language
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 65
use or processing (for example non-verbal categorization or memory tasks as in
Bohnemeyer et al. 2001; Choi and Hattrup 2012; Gennari et al. 2002; Papafragou
et al. 2002; Papafragou and Selimis 2010; Pourcel 2005; Soroli 2011; Zlatev et al.
2010) and/or non-verbal measures such as eye-tracking (Papafragou et al. 2008;
Soroli and Hickmann 2010; Soroli 2012).
A secondary methodological question concerns how to best measure non-ver-
bal performance, particularly in order to avoid the possibility that subjects might
internally verbalize information while construing event representations. For this
purpose, cognitive psychology has developed experimental paradigms using “dual
tasks” in which subjects perform a secondary concurrent task (hereaer “inter-
ference” task) while carrying out a main task (e.g. seminal work by Baddeley and
Hitch 1974). e assumption is that if a task selectively interferes with one type
of process but not with another, then those two types of processes are indepen-
dent and rely on dierent parts of the cognitive system. Dual-task paradigms can
involve material and/or behaviors that are verbal or verbal-like (e.g. shadowing
during syllable repetition) vs. non-verbal (e.g. nger tapping in a particular se-
quence) in order to interfere with dierent components of cognitive architecture.
In the context of the present paper, the rst type of interference task is the most rel-
evant (for detailed discussions see Newton and de Villiers 2007; Hermer-Vasquez
et al. 1999): “verbal shadowing” is meant to imply “articulatory suppression” pre-
venting subjects from internally verbalizing information and ensuring that they
only activate non-verbal representations that are not “contaminated” by language
(see discussion in Section4.2.4).
e present study addresses some of these questions by examining the perfor-
mance of adult native speakers of English and French in three tasks in which they
processed various types of motion events: (i) in a non-verbal categorization task
they had to make similarity judgments based only on visual information (triads
of cartoons, one target and two variants), while simultaneously performing an
interference task, (ii) in a verbal categorization task they also had to make simi-
larity judgments, but the target was a sentence and there was no interference task,
(iii) during a production task they simply described cartoons. e extent to which
language is involved in subjects’ performance across these tasks corresponds to
the following gradient: (i) language is least (and in principle not at all) involved
in the non-verbal categorization task (non-verbal stimuli, non-verbal responses,
articulatory suppression), (ii) it is partially involved in the verbal categorization
task since target sentences require subjects to process linguistic representations
and thus allow us to test the impact of verbal input on their preferential choices,
(iii) it is most involved in the production task that required event construal spe-
cically for the purpose of full-edged verbal descriptions.
2nd proofs
66 Maya Hickmann et al.
2. Methodology
2.1 Participants
Two groups of young adults participated in the study, native speakers of French
vs. English. Participants were students at the University of Cambridge (English)
and at the Universities of Paris 8 and of Lille 3 (French). A questionnaire evaluated
psycholinguistic and sociolinguistic information. Sampling excluded: bi- or mul-
tilingual subjects; those with some past or present cognitive or language problems;
those who had had extensive exposure (of more than 6 months) to another language
in another country. A total of 20 subjects were planned within each language, but
some subjects in the French sample had to be excluded due to several factors (French
N = 14, Engl ish N = 20): sociolinguistic background (four subjects), unwillingness
to nish the session (one subject), failure to respond correctly to at least half of
“control ” items in the categorization tasks (one subject, see Section3.2).
2.2 Stimuli
A set of 18 stimuli was designed (hereaer “targets”). ey consisted of black-
and-white animated cartoons
2 (720×540 pixels) showing animals (goat, bear, cat,
horse, dog, lion) performing spontaneous displacements along one of six paths
(, , ,  , , ) and in one of three manners (,
, ) against one of three backgrounds (a hill, a house, railway tracks). Paths
and manners were selected with the following rationale. First, manners showed
a continuum from least to most salient (<<). is continuum was
based on assumptions concerning which types of manners are relatively neutral or
marked for human gures, in accordance with general world knowledge regarding
human locomotion. By extension, the same scale was applied to the animal gures
used in the experiment. Second, paths fell into two main types: boundary-cross-
ings (,  , ) or motion without any boundary-crossing that was
either vertical (, ) or not (). ird, both manners and paths were
meant to be familiar to participants and as natural as possible for the gures. All
paths were crossed with all manners, resulting in 18 targets. Motion always took
place from le to right.
3 e appendix summarizes the nature of these stimuli and
illustrates one item from the non-verbal categorization task.
2. Black-and-white (grayscale) rather than colored cartoons were used in order to focus subjects’
attention only on motion events in the stimuli.
3. During previous pre-tests, motion in each cartoon occurred either from le to right or in
the reverse direction. Since the data showed no eect of direction, this variable was subsequently
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 67
Two versions of each target item were also designed as “variants” for the cat-
egorization tasks which diered from the target by virtue of Manner or Path (see
Section2.3). In addition, control and distractor items (six of each) were inter-
spersed among experimental items. Control items were identical to experimental
ones except that one variant was identical to the target (and therefore correspond-
ed to the “correct” answer, see Section2.3). ese items allowed us to control for
random performance, i.e., subjects were excluded if at least half of their responses
to these items was incorrect. Distractors showed inanimate entities in motion (e.g.
a ball rolling down a slope). ey were meant to divert subjects’ attention from the
actual goal of the study by presenting events that were not unrelated (motion) but
nonetheless quite dierent from the other items. Finally, two training items were
presented at the beginning of each task to familiarize subjects with the procedure.
Each subject saw all target stimuli and therefore all possible combinations of
Manner+Path in these stimuli, but not all possible variants for a given target. In
order not to inate the total number of triads presented to each subject, variants
diering with respect to Manner were distributed over two groups of subjects. For
example, if the target was  , one variant was   (for both Groups),
but the other variant was either   (for Group1) or  (for Group2).
2.3 Tasks and procedure
Participants were tested individually. ey were shown stimuli on a computer
screen and performed three tasks developed with E-Prime 2.0 Soware (Schneider
et al. 2002): two versions of a categorization task and a production task.
1. Non-verbal categorization
Subjects saw triads of cartoons: a target cartoon (e.g. a cat walking up a hill), im-
mediately followed by two variants that diered from it with respect to Manner
or Path (cat walking down vs. jumping up). ey had to choose which variant
best matched the target, indicating their choice by pressing a response key as fast
as possible. In contrast to control items which proposed one correct variant, no
response was correct for experimental items, the aim being to determine whether
subjects would choose Path or Manner as their preferential similarity criterion.
While making their judgments, subjects simultaneously performed an interference
task that consisted of repeating triads of syllables during the entire task. ese triads
were designed as follows: three consonants (/d/, /b/, /m/) were combined with three
vowels (/a/, /i/, /o/), then a number of these combinations were selected so that they
would be natural in both languages (e.g. “BaBiBo, MaMiMo, DaMaBa, MoDoBo”).
Triads were prerecorded by native speakers of French and English. A new syllable
triad was randomly assigned to each new item and was presented auditorily at the
2nd proofs
68 Maya Hickmann et al.
beginning of each item. For each stimulus the interference task started when the
target appeared on the screen and ended aer participants had pressed the key.
2. Verbal categorization
is task was identical to its non-verbal version above except that the target was a
sentence presented orally (rather than a cartoon presented visually) and that there
was no interference task. All target sentences were constructed so as to explicitly
encode Manner and Path. Although this is less typical for French than for English,
a natural structure expressing this combination within each language was used
for all items, e.g. (15) and (16).
(15) ere’s a cat walking up a hill
(16) Il y a un chat qui monte la colline en marchant
ere is a cat that is ascending the hill by walking’
3. Production task
Subjects were shown the cartoons that served as targets in the non-verbal catego-
rization task and they were asked to describe these cartoons verbally. Within each
task, the order of stimuli was randomized for each subject. All subjects performed
the three tasks in a xed order: rst non-verbal categorization, then verbal cate-
gorization, and nally production. e rationale for this order was that subjects
had to begin with the most non-verbal task (non-verbal categorization) so that
their responses in this task would not be “contaminated” by any task-related ver-
bal input or output. Conversely, they ended the session with the most verbal task
(production), while the intermediate task (verbal categorization) was placed in
second position (verbal input but no verbal output).
2.4 Hypotheses
In light of current debates, our main hypotheses concerned language eects
across the three tasks. It was rst predicted that production should show clear
language eects due to typological properties of French and English, specically
more manner verbs together with path elements in English and more path verbs
without or with less manner expression in French. Second, if language particulars
aect performance in non-verbal categorization, then there should be more man-
ner-choices in the English group than in the French group. Otherwise, participants
in both groups should make either equally many path- and manner-choices (no
preference) or more path-choices (if Path is most basic, irrespective of language).
ird, verbal categorization should highlight manner information since it involves
a verbal target and therefore should invite more manner-choices than non-verbal
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 69
categorization. is result was expected for both groups, but particularly for the
French participants who may be more inuenced by this variable since the target
sentence does not correspond to the most typical structure in French.
Another set of predictions concerned properties of the events (Manner and
Path) to be processed and their potential impact on performance. First, it was
predicted that the more salient the Manner (>>), the more like-
ly participants should be to pay attention to this component. Second, given the
special status of boundary-xscrossings (see Examples(13)–(14)), it was expected
that attention to Manner may be lower for boundary-crossings (,  ,
) than for items without boundary-crossing (, , ). During
production, boundary-crossings should increase path expression and/or decrease
manner expression. In a similar vein, during categorization, boundary-crossing
should elicit fewer manner-choices than other items. Although these predictions
about manner and path eects were expected to hold for both language groups,
manner eects were expected to be stronger for English participants since these
participants should focus on Manner with all events. In addition, path eects
should be stronger for French participants for whom Path is most likely to be
expressed in the verb when motion implies a boundary-crossing.
3. Results
Analyses are presented below in two sections focusing rst on subjects’ responses
during the production task (Section3.1), then on their responses in the two cate-
gorization tasks (Section3.2).
3.1 Production
e analyses of subjects’ verbal productions focus on which components of motion
they express and how they are encoded. All productions fell into three groups: Path
alone (hereaer “P-responses”, e.g. (17) and (18)); Manner alone (“M-responses”,
e.g. (19) and (20)); or both Manner and Path (“MP-responses”, e.g. (21) to (23)).
In addition, analyses identied the linguistic means used to express Manner and/
or Path, with particular attention to encoding in main verb roots vs. outside.
Expression outside the main verb included particles, prepositional phrases, and
adverbials, as well as various devices in subordinate clauses. For example, Path
4. In all analyses a probability value of up to 0.05 was taken to reect a signicant result.
5. Other aspects of productions were coded and analyzed, but cannot be discussed in detail here.
2nd proofs
70 Maya Hickmann et al.
is expressed by main verbs in (17), (22), and (23) (rentrer ‘to enter’, sortir ‘to exit’,
monter ‘to ascend’) vs. by particles or prepositions in (18) and (21) (into, up);
Manner is expressed by main verbs in (19) to (21) (galoper ‘to gallop’, run), by an
adverbial phrase in (22) (au galop, ‘by a gallop’), and by a gerund in a subordinate
clause in (23) (en courant ‘by running’).
(17) Un chat rentre [Path] dans une maison
A cat enters in[to] a house’
(18) e cat went into [Path] a hut
(19) Un cheval galope [Manner] près des rails
A horse gallops near the railroad tracks’
(20) e lion ran [Manner] next to the train tracks
(21) A bear ran [Manner] up [Path] a hill
(22) Un cheval est sorti [Path] au galop [Manner] d’une petite maison
A horse exited a little house by a gallop
(23) Un ours monte [Path] la montagne en courant [Manner]
A bear ascends the mountain by running
Figure1 shows the global distribution of all responses within each language.
Motion expressions are clearly language-specic, as reected by more frequent
MP-responses in English and more frequent P-only responses in French.
English French
Figure1. Production: Percentages of Motion components expressed in verbal descriptions
within each language (MP=both Manner and Path, P=Path only, M=Manner only)
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 71
As shown in Figure2, English responses show an almost perfect complementary
distribution in which almost all verbs encode Manner (91%) and almost all other
motion elements encode Path (90%). In contrast, the majority of French verbs
express Path (82%) without any other motion-relevant information outside of the
verb (hereaer Z, 58%), although additional elements in some responses occasion-
ally express Path (19%), Manner (15%) or both (8%).
Verb Other
Verb Other
Figure2. Production: Distribution of semantic information in verb roots and other
elements within each language (in percentages; MP=both Manner and Path, P=Path
only, M=Manner only; Z=neither Manner nor Path)
2nd proofs
72 Maya Hickmann et al.
A mixed-model ANOVA was carried out to test the effects of Language
( between-subject variable), and of path and manner types in the stimuli
( within-subject variables) on the simultaneous verbal expression of Manner and
Path (hereaer “MP-responses”). is dependent variable was chosen since our
main hypothesis concerning this task predicted more manner information in
English than in French, but did not predict any language eect with respect to path
expression. e ANOVA rst shows a signicant language eect conrming that
MP-responses were indeed signicantly more frequent in English (F(1,32) = 79.5537,
p = 0.0001). In addition, this language eect can be observed regardless of event
properties across stimuli, i.e., a signicant dierence (English>French) is system-
atically observed within all three manners and with all six paths.
Second, as expected, the relative salience of Manner ( <  < ) has a
signicant eect on MP-responses (F(2,64) = 12.4970, p = 0.0001). Contrasts also show
signicant dierences across all manners ( > : F(1,32) = 5.0746, p = 0,0297);
 > : F(1,32) = 8.5262, p = 0,0063);  > : F(1,32) = 23.0226, p = 0,0001).
However, Manner also significantly interacts with Language (F(2,64) = 3.7155,
p = 0.0292). As shown in Figure3, Manner has a strong signicant impact on MP-
responses in French (F(2,28) = 9.5367, p = 0.0007), while this eect is only marginally
signicant in English (F(2,36) = 3.1348, p = 0.0545). In particular, French speakers
are most sensitive to the relative salience of Manner, as shown by signicant con-
trasts among all manners in this sample ( > : F(1,14)=5.6875, p = 0.0305;
walk run jump
Figure3. Production: Mean MP-responses as a function of manner type in stimuli
within each language
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 73
 > : F(1,14) = 5.5547, p = 0.0321;  > : F(1,14) = 14.6638, p = 0.0019).
In comparison, English speakers produce frequent MP-responses with all stimuli
types, showing only one signicant dierence between the least salient and the most
salient manners ( > : F(1,18) = 9.075, p = 0.0073).
Path has no overall signicant eect on MP-responses, but interacts sig-
nicantly with Language (F(5,160) = 2.3526, p = 0.0428). In particular, French
participants show no signicant eect of Path at all, while Path has a signi-
cant eect on MP-responses in English (F(5,90) = 2.6974, p = 0.0258). As shown
in Figure4, although MP-responses are very frequent with all paths in English,
they are less frequent with  than with any other path (signicantly so -
> (F(1,18) = 4.4844, p = 0.0463; > (F(1,18) = 5.1634, p = 0.034;
and > (F(1,18) = 4.5685, p = 0.0446). As illustrated in Examples(24)
and (25), English descriptions of -items sometimes comprised a manner
verb with a locative expression or a “lighter” verb (to go) with a path particle.
(24) e dog walked next to the train tracks
(25) A lion went down a railway track
up down into out of across along
Figure4. Production: Mean MP-responses as a function of path types in stimuli within
each language
Finally, path and manner types interact but in dierent ways across language
groups, as shown by the signicant interaction between Path, Manner and Lan-
guage (F(10,320) = 2.0362, p = 0.0299). Figure5 shows MP-responses as a func-
tion of both event properties within each language. Beyond the higher level
2nd proofs
74 Maya Hickmann et al.
of MP-responses with all event types in English than in French, the following
main language dierences can be highlighted in these complex data, all in line
with the results discussed above. English MP-responses are least frequent with
, regardless of manner type and notwithstanding some variation with
manner salience. ey are more likely with boundary-crossings (,  ,
) when Manner is salient, but this is not the case with vertical motion (,
). In French, Manner is least expressed with  and most w ith , but
up down into out of across along
up down into out of across along
Figure5. Production: Mean MP-responses as a function of Manner and Path in stimuli
within each language group
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 75
MP-responses generally vary with manner salience, being particularly frequent
with  and/or with  as compared to , regardless of path types. us,
in both language groups speakers’ focus on Manner depends on both path and
manner properties of the events to be described.
3.2 Categorization
Recall that in the categorization task participants had to choose which of two
variant cartoons were most similar to a target cartoon (in the non-verbal condi-
tion) or corresponded best to a target sentence (in the verbal condition). Although
there was no correct answer on any of the experimental items (similarity could
be based on Manner or Path), control items did provide a correct answer (i.e., one
variant was identical to the target) in order to control for random performance.
Accordingly, subjects who had failed to respond correctly to at least half of the
control items were excluded from all subsequent analyses (one subject in French).
A mixed-model ANOVA was then performed to test the effects of Language
( between-subject variable), as well as Condition (non-verbal vs. verbal), path and
manner types (within-subject variables) on choice of Manner as criterion for sim-
ilarity judgments (hereaer “M-choices”).
During categorization, there was no overall signicant language eect on
M-choices: both language groups chose Path more frequently than Manner.
However, Condition had an overall signicant eect (F(1,32) = 6.3898, p = 0.0159).
As shown in Figure6, regardless of their language, when subjects did make
M-choices, they did so more oen in the non-verbal condition than in the verbal
condition. Furthermore, a signicant language eect is observed but only in the
verbal condition (F(1,32) = 4.799952, p = 0.0342), where M-choices were more fre-
quent in the English group than in the French group.
6. Two additional secondary factors were included in a preliminary design: the particular an-
imal Figure in the stimuli and subjects’ gender. Both variables were excluded from the analy-
ses presented below since they showed either no eect or sporadic and/or marginal eects on
2nd proofs
76 Maya Hickmann et al.
Non-verbal Verbal
Figure6. Categorization: Mean M-choices as a function of condition
within each language
Event properties also aected subjects’ performance. Manner type had a signif-
icant overall eect on their choices (F(2,64) = 5.1203, p = 0.0087). Further con-
trasts show two results. First, when collapsing languages,  (most salient
manner) elicited more M-choices overall than  (F(1,32) = 9.1605, p = 0.0048).
Second, however, there was also a signicant interaction between this variable
and Language (F(2,64) = 7.7 59 4, p = 0.001), as shown in Figure7. Manner type
signicantly inuenced M-choices in the English group (>: F(1,19) = 15,
p = 0.001; > F(1,19) = 9.2178, p = 0.0066), but did not have any signicant
eect in the French group.
Path type did not signicantly aect responses overall, but this variable sig-
nicantly interacted with Language (F(5,160) = 2.3526, p = 0.0428). As shown in
Figure8, M-choices did not vary signicantly as a function of Path in the French
group, but they did in the English group, where  and to a lesser extent 
elicited more M-choices than other paths, particularly in comparison to vertical mo-
tion (and signicantly so for >: F(1,19) = 6.7205, p = 0.0171; >:
F(1,19) = 6.5376, p = 0.0185; >: F(1,19) = 4.5398, p = 0.0445).
ere were no signicant interactions between Condition and either manner or
path type. However, a glance at the distribution of M-choices as a function of event
properties in each Condition and Language shows interesting dierences. First,
Figure9 shows this distribution as a function of each manner (collapsing paths).
Manner has no signicant eect on the French participants’ responses in either
condition, while it signicantly inuences the English group in both conditions
(non-verbal: F(2,38) = 4.7965, p = 0.0138; verbal: F(2,38) = 7. 60 68, p = 0.0017).
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 77
Note that the French group never chooses Manner as criterion with , further
suggesting that they are not sensitive to Manner. Contrasts in the English group
show that signicant dierences occurred in the non-verbal condition (>:
F(1,19) = 7.11 97, p = 0.0146; >: F(1,19) = 6.9091, p = 0.0159) and in the
English French
Figure7. Categorization: Mean M-choices as a function of manner type
within each language
out of
Figure8. Categorization: Mean M-choices as a function of path type
within each language
2nd proofs
78 Maya Hickmann et al.
verbal condition (>: F(1,19) = 13.3601, p = 0.0017), other contrasts being
marginally signicant (>: p = 0.0528, >: p = 0.0506).
Second, Figure10 shows the distribution of M-choices as a function of paths
(collapsing manners) in each condition and language. No contrast is signicant
for the French sample, regardless of condition. However, in the English sample
M-choices vary a great deal as a function of Path and more so in the verbal than
in the non-verbal condition. us, during verbal categorization English speakers
make more M-choices with  than with other paths (signicant >:
F(1,19) = 6.3333, p = 0.0201; >: F(1,19) = 5.9421, p = 0.0237;
English French
Verbal categorization
English French
Non-verbal categorization
Figure9. Categorization: Mean M-choices as a function of manner type
within each condition and language
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 79
>: F(1,19) = 6.3333, p = 0.0201; > : F(1,19) = 7. 307 7,
p = 0.0136). -stimuli also elicit more M-choices, although to a lesser extent
(signicantly so for > : F(1,19) = 5.5161, p = 0.0285, but also mar-
ginally signicant for >: F(1,19) = 4.1707, p = 0.0528, >:
F(1,19) = 4.1304, p = 0.0538, and >: F(1,19) = 4.1707, p = 0.0528).
During non-verbal categorization,  and   elicit more M-choices than
vertical motion (signicantly so for >: F(1,19) = 5.5161, p = 0.0285 and
 >: F(1,19) = 5.6296, p = 0.0272).
English French
Verbal categorization
out of
English French
Non-verbal categorization
out of
Figure10. Categorization: M-choices as a function of path type
within each condition and language (in means)
2nd proofs
80 Maya Hickmann et al.
4. Discussion
is study aimed to assess language eects in three tasks that varied in the degree
to which they involved language: (i) a highly non-verbal task (non-verbal cate-
gorization), in which participants processed non-verbal information (cartoons)
and produced non-verbal responses (similarity judgments) while performing
an interference task (syllable repetition), (ii) a highly verbal task (production) in
which they produced explicit verbal descriptions, (iii) an intermediate task (verbal
categorization) that involved non-verbal responses but presented subjects with
some verbal input to be processed (target sentences expressing Manner and Path,
followed by two variant cartoons). We rst briey summarize the main results
before proposing a discussion and concluding remarks raising some questions to
be pursued in future research.
4.1 Summary of results
Production shows massive language eects. First, regardless of event properties,
English speakers produce semantically richer responses encoding both Manner (in
verb roots) and Path (outside of the verb), while French speakers most frequently
express Path alone (in verbs). Second, the more salient the Manner, the more likely
French speakers are to add this component to Path, whereas their MP-responses
do not vary as much with path type. In comparison, English MP-responses are
generally frequent with all event types, although they are less frequent with 
stimuli than with other path types, regardless of manner type, and they are more
likely with boundary-crossings (,  , ) when Manner is salient
Categorization shows fewer language eects and more complex results. First,
both groups choose Path more frequently than Manner as categorization criterion.
Second, when speakers (of either group) do make M-choices, they do so more oen
in the non-verbal condition than in the verbal condition. ird, event properties
signicantly aect subjects’ choices in the English group but not in the French
group. For English participants M-choices are most frequent with  (most
salient manner) but also surprisingly, more frequent with  (least salient) than
with  (intermediate). eir M-choices also vary as a function of Path, being
most frequent with  and least frequent with /. Sensitivity to event
properties in the English group is observed in both verbal and non-verbal con-
ditions, but more so in the verbal condition, notwithstanding the overall greater
frequency of M-choices in the non-verbal condition.
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 81
4.2 Implications
e discussion below rst returns to our main research question focusing on lan-
guage eects as a way to test the impact of linguistic relativity both on verbal
production and beyond. We then turn to other questions that emerge from our
study: the impact of event properties on subjects’ performance, the implications
of some observed variation, as well as methodological issues concerning how to
tap subjects’ mental representations.
4.2.1 Verbal and non-verbal cognition
As predicted, participants do not verbalize the same motion information in the
production task as a result of the particular properties of their language. is result
replicates the ndings of many previous studies showing systematic crosslinguistic
variation in motion expression by adults and children from the youngest ages on
(Allen et al. 2007; Choi 2011; Choi and Bowerman 1991; Harr 2012; Hickmann et
al. 2009a, b; Ji 2009; Ji et al. 2011a, b; Slobin 1996, 2004). However, current debates
have highlighted the necessity to go beyond language production in order to show
language eects on our mental representations, particularly in order to avoid the
circularity that would be otherwise inevitable if we merely claimed that language
particulars inuence language use. For unconvinced researchers, such a statement
would amount to saying that “language inuences language”. One way out of this
circularity is to compare the performance of speakers from dierent language
groups in verbal and non-verbal tasks. e present study compared three tasks
that varied in the degree to which subjects were “contaminated” by language. We
discuss the results below by comparing rst production vs. categorization, then
non-verbal vs. verbal categorization. Production vs. categorization
One main conclusion of our study is that language particulars massively aect pro-
duction, whereas they do not signicantly aect categorization overall. is result
suggests that language particulars aect cognition when the task is most verbal
but not when it is most non-verbal. However, before reaching this conclusion, it
is necessary to account for signicant language eects related to event properties
across tasks. In particular, manner and path properties have dierent eects on
speakers’ verbal descriptions across the two language groups during production
and they only aect responses in the English group during categorization.
In the French group MP-responses in production are only frequent when
Manner is salient (>>) and they do not vary with Path. In com-
parison, they are least frequent in the English group with , but do not vary
with Manner. is rst set of results reects speakers’ preferred structures when
2nd proofs
82 Maya Hickmann et al.
describing motion, which follow directly from their typological properties: MP-
responses are most frequent in English, where speakers represent Manner verbally
regardless of the manner to be described; in contrast, French speakers most fre-
quently express Path only, regardless of the path to be described, and they only
add Manner when this component is salient.
Turning to categorization, M-choices vary signicantly as a function of event
properties, but only in the English group. Whereas M-choices in the French group
are indierent to stimulus type, English participants pay attention to both man-
ner and path properties when categorizing events. M-choices are most frequent
when Manner is highly salient () but also very frequent when it is least salient
(). In addition, although not all path contrasts are signicant, the general
pattern emerging for the English sample shows the highest rate of M-choices with
 and the lowest with vertical motion (/), while boundary-crossing
events occupy an intermediate position (,  , ). ese results hold
across categorization conditions, notwithstanding the overall greater frequency
of M-choices in the non-verbal condition and the larger number of signicant
dierences across event types for the English group in the verbal condition.
Before we return to event properties in more detail below, note that such lan-
guage eects, specically related to particular event properties in dierent tasks,
show that it is necessary to go beyond the presence/absence of overall signicant
language eects and to look more closely into the conditions under which such
eects are most likely to emerge as a function of the task to be performed and of the
properties (both Manner and Path) of the events to be processed (see also Slobin
2006). It seems that binary distinctions (such as S- vs. V-languages) are insucient
to predict categorization preferences and that studies of event types taking into
account more ne-grained distinctions are necessary to address these questions. Verbal vs. non-verbal categorization
Predictions concerning language eects in the categorization tasks were as follows:
(i) if categorization should show a language eect, then the English group should
focus more on Manner than the French group (M-choices), (ii) otherwise, choices
in both groups should show either no preference (M and P) or a greater focus on
Path, (iii) the verbal categorization condition should show more M-choices than
the non-verbal condition, particularly in the French group.
Our results rst show no overall signicant language eect in participants’
choices, but rather a clear preference for Path as categorization criterion in both
groups. According to our initial hypothesis, this result might suggest that Path
constitutes the most basic component dening motion, irrespective of language
particulars. In line with this interpretation, note that (static) ground information
in our visual stimuli (hill, house, railroad tracks) was immediately available as
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 83
soon as it appeared on the screen. erefore, a global glance at the stimuli may
be sucient to infer at least some paths. In contrast, judgments about (dynamic)
manner information is based on a myriad of details (e.g. the use of dierent body
parts, changes in body position, the amplitude and speed of body movements)
and probably require motion to start unfolding before it can be fully processed
(particularly in the absence of obvious instruments such as bicycles). As a result,
subjects could have chosen Path as their preferred criterion, irrespective of lan-
guage group, especially given task constraints (responding as fast as possible while
simultaneously performing an interference task).
Second, ndings also show more M-choices in the non-verbal categorization
task than in the verbal condition in both language groups. Contrary to expec-
tation, if we look at the overall data, verbal input (target sentence) did not in-
duce a signicantly greater focus on Manner than visual input (target cartoon).
ird, however, a language eect was observed in verbal categorization (and not
in non-verbal categorization), suggesting that verbal input did have an eect on
categorization but only in the English group. us, whereas M-choices did not
dier across language groups in the non-verbal condition, they were signicantly
more frequent among English participants than among French ones in the verbal
condition. is language eect could be due to the fact that English target sen-
tences (manner verb followed by path element) lexicalized Manner in the verb,
i.e., highlighted Manner early in the sentence and in a structure that was typical
for this language. In contrast, French speakers heard sentences in which Path was
lexicalized in the verb and Manner expressed peripherally. us, although path
verbs are typical in French, both the explicit expression of Manner (less typical in
French) and its position (late in the sentence) probably made it less likely for these
participants to focus on Manner in the verbal condition.
Finally, as mentioned above, event properties (Manner and Path) aect sub-
jects’ categorization performance, but only in the English group, and they do so
more in the verbal condition than in the non-verbal condition. is result suggests
again that verbal input did aect information processing during categorization in
the English group, but not in the French group. In order to understand this eect,
we must turn to a more detailed discussion of event properties.
4.2.2 Event properties
Our ndings indicate that dierent manners and paths across events impact sub-
jects’ production and categorization preferences. ese ndings partially support
our hypotheses concerning event properties, but they also raise some questions.
It was predicted that attention to Manner should vary depending on the relative
salience of this component in the stimuli, i.e., the more salient the Manner, the
more it should attract attention (>>). In addition, performance was
2nd proofs
84 Maya Hickmann et al.
expected to be inuenced by path properties, i.e., boundary-crossing (, 
, ) should invite participants to focus less on Manner than motion with-
out boundary-crossing (, , ).
In production, the prediction concerning Manner primarily holds for the
French group, where participants mainly focus on Path, adding Manner only
when it is salient. In contrast, English speakers systematically encode Manner
in the verb, leaving the periphery free to systematically add path information,
notwithstanding some inuence of manner salience in some cases ( and
boundary-crossings). Both languages show a complex interaction between Manner
and Path, such that the likelihood of MP-responses depends on both motion com-
ponents. As a result, it is not the case that boundary-crossings per se invite less
attention to Manner than other path types.
In categorization, although both groups choose Path more oen than Manner
as criterion, event properties aect English participants’ choices. In this group
M-choices are most frequent with  but also surprisingly frequent with ,
reecting a greater sensitivity to all manners on the part of English participants.
In addition, M-choices vary with Path according to the following general pattern
>,  , >/.
In order to account for this general pattern during categorization, and con-
sidering again the relative accessibility of dierent motion components in our
cartoons, note that ground information was more accessible with some paths than
with others. e Ground for / motion (hill) could have immediately indi-
cated to participants that they should expect vertical motion. e Figure’s initial
position gave them a further clue as to the direction of motion (on top for down-
ward motion, on bottom for upward motion). Similarly, with boundary-crossing
the Figure was initially positioned as follows: at the starting point for an exit
(inside house); going in the direction of the goal for an entry (towards house);
perpendicular to the Ground for a crossing event (railroad tracks). In all cases,
a glance at this initial position in relation to the Ground may have facilitated
participants’ processing of Path. In sharp contrast, with -items the Figure
appeared on the side of the Ground and parallel to it, resulting in a much more
indeterminate Path. A-items could therefore have made it more dicult for
subjects to process motion information, particularly to anticipate trajectories. If
subjects hesitated more in determining Path in these cases, they presumably were
more likely to choose Manner as their categorization criterion in comparison to
other items. is hypothesis applies particularly to participants in the English
group, who are more inclined to pay attention to Manner in comparison to French
speakers. Support for this hypothesis can also be found in the production data,
which show signicantly fewer MP-responses with -items in English (most-
ly replaced by M-only responses, e.g. walk next to in Example(24)). In addition, as
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 85
mentioned above, the verbal condition implied structures that highlighted Manner
and were quite typical for English, leading participants in this group to focus more
on Manner when responding to the cartoon variants.
Eects due to event properties suggest the need to rene the general distinc-
tion between categorical vs. gradual changes of location and to further dierentiate
events within each of these types, e.g., gradual changes of location that do vs. do
not imply a vertical axis (/ vs. ). In this respect, it is possible that
vertical motion has a special status resulting from particular aspects of embodied
cognition, e.g., a greater focus on Path or a lesser focus on Manner due to our
sense of gravity in relation to our body. Note, however, that our results do not
generally allow us to determine whether the impact of event properties is based
on path salience or on manner salience, i.e., whether some events should invite
participants to focus more on Path and/or less on Manner in comparison to other
events. Related results can be found in Zlatev et al. (2010) who argue that both
language-independent and language-specic factors account for relative focus on
Path or Manner.
4.2.3 “Degrees of freedom” and variation
Further points to be noted concern the “degrees of freedom” from which partic-
ipants benetted in all tasks. First, verbal descriptions showed dierent levels of
variation in the two language groups. English productions showed very frequent
manner expression in MP-responses with an almost perfect complementary dis-
tribution between manner verbs and path elements outside of the verb. In sharp
contrast to this highly systematic pattern, although French speakers typically ex-
pressed Path in verbs without manner information, responses were more variable
in this group.
us, some French verbs occasionally encoded both Manner-and-Path. For
example, in (26) the verb escalader ‘to climb.up’ necessarily implies the use of
specic limbs to move in an upwards direction as well as some mountain-like
ground and/or some diculty carrying out motion. Second, some manner verbs
occurred, mostly in descriptions of -items as shown in (27) and (28), and
more rarely in response to - and -items as in (29) and (30).
(26) Un chat escalade [Manner+Path] la montagne
A cat is climbing up the mountain
(27) Un cheval saute [Manner] le long [Path] d’une voie ferrée
A horse is jumping along railroad tracks
(28) C’est un lion qui court [Manner] à côté [Location] des rails
‘It’s a lion that is running next to the railroad tracks
2nd proofs
86 Maya Hickmann et al.
(29) Un chat sautille [Manner] sur [Location] la colline
A cat is hopping on the hill’
(30) Un ours qui sautille [Manner] dans [Location] une maison
A bear that is hopping in[into] a house’
Furthermore, French elements outside of the verb most frequently denoted gen-
eral locations, as well as sometimes either Manner or Path. Prepositional phrases
in (28) and (29) above (à côté des rails ‘next to railroad tracks, sur la colline ‘on
the hill’) clearly express general locations; (30) (dans une maison ‘in a house’)
also strongly invites a reading whereby the prepositional phrase denotes a gen-
eral location, notwithstanding some potential ambiguity. Examples such as (31)
(jusqu’en haut d’une colline ‘all the way to the top of a hill’) show more typical
uses of prepositional phrases expressing path information with upwards motion,
namely the endpoint of motion. Finally, subordinate clauses were used in several
ways. Most frequently, gerunds added manner information to a main path verb as
in Example(23) above (monter en courant ‘to ascend by running’). Occasionally,
however, gerunds also added Path to a main manner verb as in (32) (sauter en
longeant ‘to jump going.along’) and some innitival purpose clauses added Path
to a main manner verb as in (33) (sautiller pour sortir ‘to hop in order to go out’).
(31) Un chat sautille [Manner] jusqu’en haut [Path] d’une colline
A cat hops all the way to the top of a hill’
(32) Un cheval saute [Manner] en longeant [Path] un chemin de fer
A horse is jumping while going along railroad tracks
(33) Un chien qui sautille [Manner] pour sortir [Path] de la maison
A dog that hops in order to exit the house
Second, a more general point about variation should be kept in mind. Recall that
none of our three tasks involved “correct” answers. In both categorization tasks,
subjects could preferentially rely on Manner or on Path as possible criteria to clas-
sify motion events. Similarly, in the production task, they could focus on dierent
aspects of motion, using a range of structures available in their language to do
so, all of which are perfectly grammatical, but only some of which correspond to
typical language-specic patterns. Consequently, it is important to stress that sub-
jects’ responses were not obligatorily “dictated” by task constraints nor by gram-
mar-specic constraints, but rather by typical structures in their language. Such
degrees of freedom argue against the idea that the linguistic relativity hypothesis
is “circular”, i.e., it is not merely the case that “language inuences language” in
an entirely deterministic way since subjects always have a choice.
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 87
4.2.4 Methodological issues
A nal set of methodological questions must be raised for the sake of future re-
search. A rst point concerns possible artefacts due to our stimuli. For example,
recall that participants’ attention to manner properties was greater in the English
sample, including with  which was meant to be the least salient manner. We
interpret these ndings as being consistent with the prediction that all manner
information should be highly salient for English speakers. However, it is also not
impossible that frequent M-choices with  may be due to the fact that bodily
movements were not suciently precise and/or typical for the animal gures in
our stimuli, perhaps resulting in an unexpectedly high salience of  and/or
in some diculties for subjects when processing these items. Indeed, subjects
occasionally expressed some hesitation as to manner type in the production task
(e.g. He was walking in a strange way). Although this possibility cannot be entirely
excluded, other aspects of our data do not support it, e.g., in their descriptions
participants predominantly used the manner expressions that were expected on
the basis of the stimuli.
Second, when processing motion events, we rely on varied types of informa-
tion that serve as the basis for our judgments about Manner and Path, raising more
general questions about the types of stimuli that are best suited to test sensitivity
to these event properties. Although our stimuli allow a high level of control over
many variables, they may not be as “ecological” as other types of stimuli, such
as lms of real gures (including humans) performing real motion in real set-
tings. Yet deeper questions arise concerning how to best dene Manner or Path,
irrespective of stimuli type (cartoons, videos), in order to operationalize them in
experimental situations. Confronting this question has led some researchers to
simplify event properties at the expense of ecological validity and running the
risk of proposing partially misleading conclusions about the processing of Manner
and Path. For example, authors frequently reduce Manner to body movements that
are easy to distinguish but are quite articial (e.g. using gures without limbs,
Bohnemeyer et al. 2001), or by reducing Path to only one of its components, e.g.,
equating it with endpoints (Papafragou et al. 2008) which are clearly insucient
by themselves as criterial for Path.
ird, the order in which tasks are presented should be counterbalanced in
future research, given that subjects saw the same stimuli across tasks. As explained
above (see Section2.3), a xed task order was clearly necessary in the present study
(non-verbal categorization, verbal categorization, production) so that participants’
performance in the most non-verbal task would be least “contaminated” by any
language input or output. However, this order resulted in participants’ unbalanced
knowledge of the stimuli, particularly across the two categorization conditions.
In addition, the verbal categorization task required intermodal processing, i.e.
2nd proofs
88 Maya Hickmann et al.
mapping verbal input onto visual information. Alternative ways of contaminating
performance include presenting the production task rst, i.e., in order to deter-
mine whether subjects’ own productions might contaminate their categorization
Fourth, additional on-line measures provide interesting complementary in-
formation. For example, measuring how fast subjects respond informs us about
the relative ease or diculty with which they process motion information.
7 In
addition, eye-tracking data in all three tasks would also allow us to measure where
subjects allocate their attention when preparing to perform dierent tasks (verbal
descriptions vs. categorization with or without verbal input). One of our studies in
progress presently uses video stimuli that show real human gures moving against
the backdrop of real scenes, a new verbal categorization task, and a new memory
task, coupling all tasks with reaction time and eye-tracking measures (also see
Engemann et al. 2015; Hickmann et al. 2014; Soroli 2011; Soroli and Hickmann
2010; Soroli et al. 2015).
Finally, it is time for future research to reevaluate the implications of dual tasks
for data interpretation. Previous studies have simply taken interference tasks for
granted as a means of ensuring articulatory suppression without seriously weigh-
ing the nature of these tasks (cf. Newton and de Villiers 2007; also see Lavie et al.
2004 for a discussion on “load theory of attention”). For example, it is not clear
how “linguistic” the material must be (e.g. syllables vs. words). In addition, little is
known about the level of complexity that is required or possible during dual tasks.
Interference tasks must not be too complex (in order to avoid too much disruption
in performance), but they must also not be too simple (in order to ensure articula-
tory suppression). Syllable repetition has been typically used with little concern for
the possible eects of many factors (e.g. precise phonological properties of sylla-
bles, number of syllables to be repeated at one time, same or dierent syllables with
each new item) and for the quality of subjects’ repetitions (e.g. errors, hesitations,
blanks; see Choi and Hattrup 2012). Finally, a more general meta-methodological
question remains open: it is still unclear whether, how, and why articulatory sup-
pression should really guarantee non-verbal performance during the main task,
for example in comparison to non-verbal categorization without any interference.
Data concerning levels of brain activation (Zhou et al. 2006) as a function of the
material and task would be of interest to address such questions.
7. Although they could not be presented because of space limitations, reaction times were
recorded in our study and these data partially support some of our hypotheses above, e.g. longer
reaction times for -items in the English group.
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 89
5. Concluding remarks
ese results point to two preliminary conclusions. First, language particulars
inuence verbal cognition more than non-verbal cognition. Language-specic
patterns mostly occur in the production task but less so in either categorization
tasks (verbal and non-verbal). Second, however, Manner and Path components are
not accessible to the same extent to speakers of typologically dierent languages
and the accessibility of these event properties further varies as a function of event
types and tasks. When producing verbal descriptions, English speakers frequently
mention both components, while French speakers frequently focus on Path alone,
suggesting that Manner is less accessible to them in production. During categori-
zation, Path is generally more basic and accessible than Manner for both language
groups. However, manner-choices are overall more frequent in the non-verbal
condition as well as more frequent for participants in the English group who are
more sensitive to both manner and path properties than French speakers, partic-
ularly when categorization involves some verbal input.
us, although our study shows more language eects in tasks involving lan-
guage use or processing, it also reveals subtle and complex language eects in
non-verbal tasks. Such eects were not reported to this degree of detail by previous
studies. ey do not support the claim that language particulars have no inu-
ence at all on non-verbal performance. Instead, they reveal systematic variation
in performance as a function of both manner and path properties across language
groups. Such language eects clearly require further research that must take into
account a number of methodological issues in order to test the impact of language
particulars on non-verbal cognition.
is research is part of the project “LANGACROSS2” nanced by the French and German na-
tional funding agencies (ANR, DFG).
Allen, S., Özyürek, A., Kita, S., Brown, A., Furman, R., & Ishizuka, T. 2007. Language-specic and
universal inuences in childrens syntactic packaging of Manner and Path: A comparison of
English, Japanese, and Turkish. Cognition, 102(1), 16–48. doi: 10.1016/j.cognition.2005.12.006
Aske, J. 1989. Path predicates in English and Spanish: A closer look. Proceedings of the Fieenth
Annual Meeting of the Berkeley Linguistics Society, 15, 1–14. doi: 10.3765/bls.v15i0.1753
2nd proofs
90 Maya Hickmann et al.
Baddeley, A. D., & Hitch, G. 1974. Working memory. In G. H. Bower (Ed.), e psychology of
learning and motivation: Advances in research and theory. Vol. 8 (47–89). New York: Aca-
demic Press.
Blomberg, J., & Zlatev, J. 2009. Linguistic relativity, mediation and the categorization of motion.
In J. Zlatev, M. Andrén, M. Johansson Falck, & C. Lundmark (Eds.), Studies in language and
cognition (46–61). Newcastle upon Tyne: Cambridge Scholars Publishing.
Bohnemeyer, J., Eisenbeiss, S., & Narasimhan, B. 2001. Event triads. In S. C. Levinson, & N.Eneld
(Eds.), ‘Manual’ for the eld season 2001 (100–114). Nijmegen: Max Planck Institute for
Bowerman, M., & Choi, S. 2003. Space under construction: Language-specic categorization in
rst language acquisition. In D. Gentner, & S. Goldin-Meadow (Eds.), Language in mind:
Advances in the study of language and thought (387–427). Cambridge, MA: MIT Press.
Cadierno, T. 2004. Expressing motion events in a second language: A cognitive typological per-
spective. In M. Achard, & S. Niemeyer (Eds.), Cognitive linguistics, second language acqui-
sition, and foreign language teaching (13–49). Berlin: Mouton de Gruyter. doi: 10.1515/978
Cadierno, T. 2008. Learning to talk about motion in a foreign language. In P. Robinson, & N. C.
Ellis (Eds.), Handbook of cognitive linguistics and second language acquisition (239–274).
London: Routledge.
Cadierno, T., & Lund, K. 2004. Cognitive linguistics and second language acquisition: Motion
events in a typological framework. In B. VanPatten, J. Williams, S. Rott, & M. Overstreet
(Eds.), Form-meaning connections in second language acquisition (139–154). Mahwah, NJ:
Lawrence Erlbaum Associates.
Carroll, M., Lambert, M., Weimar, K., Flecken, M., & von Stutterheim, C. 2012. Tracing trajecto-
ries: Motion event construal by advanced L2 French-English and L2 French-German speak-
ers. Language, Interaction and Acquisition (LIA), 3(2), 202–230. doi: 10.1075/lia.3.2.03car
Choi, S. 2011. Language-specicity of motion event expressions in young Korean children.
Language, Interaction and Acquisition (LIA), 2(1), 157–184. doi: 10.1075/lia.2.1.07cho
Choi, S., & Bowerman, M. 1991. Learning to express motion events in English and Korean: e
inuence of language-specic lexicalization patterns. Cognition, 41(1–3), 83–121.
doi: 10.1016/0010-0277(91)90033-Z
Choi, S., & Hattrup, K. 2012. Relative contribution of cognition/perception and language on
spatial categorization. Cognitive Science, 36, 102–129. doi: 10.1111/j.1551-6709.2011.01201.x
Engemann, H. 2012. Motion event expression in bilingual first language acquisition. Ph.D.
Dissertation, University of Cambridge.
Engemann, H., Hendriks, H., Hickmann, M., Soroli, E., & Vincent, C. 2015. How language im-
pacts memory of motion events in English and French. Cognitive ProcessingInternational
Quarterly of Cognitive Science 16(supplement 1), 209–213. doi: 10.1007/s10339-015-0696-7
Evans, N., & Levinson, S. C. 2009. e myth of language universals: Language diversity and its
importance for cognitive science. Behavioral and Brain Sciences, 32(5), 429–448.
doi: 10.1017/S0140525X0999094X
Filipović, L. 2011. Speaking and remembering in one or two languages: Bilingual vs. monolingual
lexicalization in memory for motion events. International Journal of Bilingualism, 15(4),
466–485. doi: 10.1177/1367006911403062
Gennari, S. P. , S l o m a n , S . A., Malt, B. C., & Fitch, W. T. 2002. Motion events in language and
cognition. Cognition, 83, 49–79. doi: 10.1016/S0010-0277(01)00166-4
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 91
Gumperz, J. J., & Levinson, S. C. 1996. Rethinking linguistic relativity. Cambridge: Cambridge
University Press.
Harr, A.-K. 2012. e impact of language-specic factors in rst language acquisition: e expres-
sion of motion in French and German. Berlin: Mouton de Gruyter. doi: 10.1515/9781614511748
Hendriks, H., & Hickmann, M. 2011. Expressing voluntary motion in a second language: English
learners of French. In V. Cook, & B. Bassetti (Eds.), Language and bilingual cognition (315–
339). New York: Psychology Press.
Hendriks, H., Hickmann, M., & Demagny, A.-C. 2008. How English native speakers learn to
express caused motion in English and French. Acquisition et Interaction en Langue Étrangère,
27, 15–41.
Hermer-Vazquez, L., Spelke, E. S., & Katsnelson, A. S. 1999. Sources of exibility in human cog-
nition: Dual task studies of space and language. Cognitive Psychology, 39, 3–36.
doi: 10.1006/cogp.1998.0713
Hickmann, M. 2010. Linguistic relativity in rst language acquisition: Spatial language and cogni-
tion. In M. Kail, & M. Hickmann (Eds.), Language acquisition across linguistic and cognitive
systems (125–146). Amsterdam: John Benjamins. doi: 10.1075/lald.52.10hic
Hickmann, M., & Robert, S. 2006. Space in languages: Linguistic systems and cognitive categories.
Amsterdam: John Benjamins. doi: 10.1075/tsl.66
Hickmann, M., Hendriks, H., & Champaud, C. 2009a. Typological constraints on motion
in French and English child language. In J. Guo, E. Lieven, S. Ervin-Tripp, N. Budwig,
K.Nakamura, & . Özçalıkan (Eds.), Crosslinguistic approaches to the psychology of language:
Research in the tradition of Dan Isaac Slobin (209–224). New York: Psychology Press.
Hickmann, M., Taranne, P., & Bonnet, P. 2009b. Motion in rst language acquisition: Manner
and Path in French and English child language. Journal of Child Language, 36(4), 705–741.
doi: 10.1017/S0305000908009215
Hickmann, M., Soroli, E., Engemann, H., Hendriks, H., & Vincent, C. 2014. Typological factors
in the development of verbal and non-verbal spatial cognition: A comparison of French
and English. Paper presented at the 5th International Conference in Cognitive Linguistics,
University of Lancaster, Lancaster, UK, 29–31 July 2014.
Ibarretxe-Antuñano, I. 2009. Path salience in motion events. In J. Guo, E. Lieven, N. Budwig,
S. Ervin-Tripp, K. Nakamura, & S. Özçalıkan (Eds.), Crosslinguistic approaches to the psy-
chology of language: Research in the tradition of Dan Isaac Slobin (403–414). New York:
Psychology Press.
Ji, Y. 2009. e expression of voluntary and caused motion events in Chinese and in English:
Typological and developmental perspectives. Ph.D. Dissertation, University of Cambridge.
Ji, Y., Hendriks, H., & Hickmann, M. 2011a. e expression of caused motion events in Chinese
and in English: Some typological issues. Linguistics, 49(5), 1041–1076.
doi: 10.1515/ling.2011.029
Ji, Y., Hendriks, H., & Hickmann, M. 2011b. Childrens expression of voluntary motion events in
English and Chinese. Journal of Foreign Languages, 34(4), 2–20.
Lavie, N., Hirst, A., de Fockert, J., & Viding, E. 2004. Load theory of selective attention and cog-
nitive control. Journal of Experimental Psychology: General, 133(3), 339–354.
doi: 10.1037/0096-3445.133.3.339
Levinson, S. C. 2003. Space in language and cognition. Cambridge: Cambridge University Press.
doi: 10.1017/CBO9780511613609
Levinson, S. C., & Wilkins, D. 2006. Grammars of space: Explorations in cognitive diversity.
Cambridge: Cambridge University Press. doi: 10.1017/CBO9780511486753
2nd proofs
92 Maya Hickmann et al.
Lucy, J. A. 1992. Language diversity and thought. Cambridge: Cambridge University Press.
doi: 10.1017/CBO9780511620843
Majid, A., Bowerman, M., Kita, S., Haun, D. B., & Levinson, S. C. 2004. Can language restructure
cognition? e case for space. Trends in Cognitive Science, 8, 108–114.
doi: 10.1016/j.tics.2004.01.003
Newton, A. M., & de Villiers, J. G. 2007. inking while talking: Adults fail nonverbal false-belief
reasoning. Psychological Science, 18, 574–579. doi: 10.1111/j.1467-9280.2007.01942.x
Özçalıkan, . 2009. Learning to talk about spatial motion in language-specic ways. In J. Guo,
E. Lieven, S. Ervin-Tripp, N. Budwig, K. Nakamura, & . Özçalıkan (Eds.), Crosslinguistic
approaches to the psychology of language: Research in the tradition of Dan Isaac Slobin (263–
276). New York: Psychology Press.
Özçalıkan, . 2015. Ways of crossing a spatial boundary in typologically distinct languages.
Applied Psycholinguistics, 36(2), 485–508. doi: 10.1017/S0142716413000325
Papafragou, A., Hulbert, J., & Trueswell, J. 2008. Does language guide event perception? Evidence
from eye movements. Cognition, 108, 155–184. doi: 10.1016/j.cognition.2008.02.007
Papafragou, A., Massey, C., & Gleitman, L. 2002. Shake, rattle, ’n’ roll: e representation of mo-
tion in language and cognition. Cognition, 84, 189–219. doi: 10.1016/S0010-0277(02)00046-X
Papafragou, A., & Selimis, S. 2010. Event categorisation and language: A cross-linguistic study
of motion. Language and Cognitive Processes, 25, 224–260. doi: 10.1080/01690960903017000
Pourcel, S. 2005. Relativism in the linguistic representation and cognitive conceptualization of mo-
tion events across verb-framed and satellite-framed languages. Ph.D. Dissertation, University
of Durham.
Schneider, W., Eschman, A., & Zuccolotto, A. 2002. E-Prime reference guide. Pittsburgh, PA:
Psychology Soware Tools Inc.
Slobin, D. I. 1996. From “thought and language” to “thinking for speaking”. In J. J. Gumperz,
& S. C. Levinson (Eds.), Rethinking linguistic relativity (70–96). Cambridge: Cambridge
University Press.
Slobin, D. I. 2004. How people move: Discourse eects of linguistic typology. In C. L. Moder, &
A. Martinovic-Zić (Eds.), Discourse across languages and cultures (195–210). Amsterdam:
John Benjamins. doi: 10.1075/slcs.68.11slo
Slobin, D. I. 2006. What makes manner of motion salient? Explorations in linguistic typology, dis-
course, and cognition. In M. Hickmann, & S. Robert (Eds.), Space across languages: Linguistic
systems and cognitive categories (59–81). Amsterdam: John Benjamins.
doi: 10.1075/tsl.66.05slo
Slobin, D. I., Ibarretxe-Antuñano, I., Kopecka, A., & Majid, A. 2014. Manners of human gait: A
crosslinguistic event-naming study. Cognitive Linguistics, 25, 701–741.
doi: 10.1515/cog-2014-0061
Soroli, E. 2011. Language and spatial cognition in French and in English: Crosslinguistic perspectives
in aphasia. Ph.D. Dissertation, University of Paris 8.
Soroli, E. 2012. Variation in spatial language and cognition: Exploring visuo-spatial thinking
and speaking cross-linguistically. Cognitive Processing– International Quarterly of Cognitive
Science, 13(1), 333–337. doi: 10.1007/s10339-012-0494-4
Soroli, E., & Hickmann, M. 2010. Language and spatial representations in French and in English:
Evidence from eye-movements. In G. Marotta, A. Lenci, L. Meini, & F. Rovai (Eds.), Space
in language (581–597). Pisa: ETS.
2nd proofs
Chapter2. Expressing and categorizing motion in French and English 93
Soroli, E., Hickmann, M., Hendriks, H., Engemann, H., & Vincent, C. 2015. Language eects on
spatial cognition? Cross-linguistic evidence and eye-tracking. Paper presented at NINJAL
International Symposium: Typology and Cognition in Motion Event Descriptions. Tokyo,
Japan, 24–25 January 2015.
Talmy, L. 2000. Toward a cognitive semantics. Vol. II: Typology and process in concept structuring.
Cambridge, MA: MIT Press.
von Stutterheim, C., & Nüse, R. 2003. Processes of conceptualisation in language production:
Language specic perspectives and event construal. Linguistics, 41, 851–881.
Zhou, X., Chen, C., Zhang, H., Xue, G., & Dong, Q. 2006. Neural substrates for forward and back-
ward recitation of numbers and the alphabet: A close examination of the role of intraparietal
sulcus and perisylvian areas. Brain Research, 1099(1), 109–120
doi: 10.1016/j.brainres.2006.01.133
Zlatev, J., Blomberg, J., & David, C. 2010. Translocation: Language and the categorization of
experience. In V. Evans (Ed.), Language, cognition, and space: e state of the art and new
directions (389–418). Sheeld, UK: Equinox.
Information presented in target stimuli
Manner Path Ground Figure
walk up hill goat
run up hill bear
jump up hill cat
walk down hill horse
run down hill dog
jump down hill lion
walk into house cat
run into house goat
jump into house bear
walk out of house lion
run out of house horse
jump out of house dog
walk across railway bear
run across railway cat
jump across railway goat
walk along railway dog
run along railway lion
jump along railway horse
2nd proofs
94 Maya Hickmann et al.
(Fixation 700 msec )
« babibo »
(continuous repetition)
(6000 msec)
6000 ms
6000 ms
3000 ms
700 ms
Example of a stimulus in the non-verbal categorization task
... Psycholinguistic studies have proposed a variety of methods to investigate these issues. Such methods include experiments using multimodal tasks, i.e. non-verbal tasks and/or co-verbal behavior that involves language to different extents, such as categorization, priming, memory tasks, often coupled with other non-verbal measures such as reaction times, gestures or eye tracking (Casasanto and Jasmin 2012;Engemann et al. 2015;Gennari et al. 2002;Hickmann et al. 2017; Soroli 2012a; Trueswell and Papafragou 2010 among others). The following sections offer an overview of such experimental studies that investigate the specificities of linguistic systems, their typological status, and most importantly the cognitive implications of language-specific features for human behavior, particularly for visual attention. ...
... More recently, Hickmann et al. (2017) compared how English and Frenchspeaking participants performed three tasks involving motion events: a non-verbal categorization task within a dual task paradigm involving articulatory suppression; a verbal categorization task involving target sentences; and a production task based on dynamic cartoon stimuli. They showed that although more language effects occur in tasks involving language use or processing, subtle and complex language effects also occur in non-verbal tasks. ...
... Recent psycholinguistic studies have proposed a variety of methods measuring the impact of language-specific variation on non-verbal spatial cognition. These methods include experiments using multimodal tasks, such as similarity judgments, priming, memory tasks often coupled with other non-verbal measures such as reaction times, gestures or eye tracking (see also Casasanto and Jasmin 2012;Engemann et al. 2015;Fibigerova and Guidetti this volume;Gennari et al. 2002;Hickmann et al. 2017;Ji and Hohenstein 2017;Trueswell and Papafragou 2010). ...
... Furthermore, for French, researchers have reported evidence of a large number of manner verb types and tokens in motion descriptions (cf. among others, Egan and Rawoens 2013;Blomberg 2014;Egan 2015;Hickmann et al. 2017; Morita this volume). However, these important nuances about French (further discussed in Section 6) do not undermine the general consensus that this language is clearly situated more towards the verb-framed pole on the full spectrum of languages compared to English or Dutch. ...
Full-text available
This study investigates visual motion expressions in Dutch, English, and French. As a translation corpus, I use Roald Dahl’s children’s book The Witches, which abounds in staring and peeping events, and its Dutch and French translations. Based on the hypothesis that languages’ constructional repertoires for physical motion are exploited for visual motion, one can predict, correctly, that Dutch uses its syntactically wide variety of path complement types in the domain of visual motion. It is tempting to assume that French, lacking looking verbs expressing path, would lose its generally verb-framed nature in visual motion descriptions. However, French appears to preserve some of its typological identity, by using causative path verbs such as lever ‘raise’ combined with an object meaning ‘one’s eyes/gaze’. In keeping with its verb-framed nature, French uses fewer visual path complements than Dutch and English, but it does have, and frequently uses, manner-of-vision expressions.
... Some studies (Choi & Hattrup, 2012;Gennari, Sloman, Malt, & Fitch, 2002;Hickmann, Engemann, Soroli, Hendriks, & Vincent, 2017;Papafragou, Hulbert, & Trueswell, 2008;Richmond, Zhao, & Burns, 2015) have begun to address these questions using other methodologies, such as a combination of verbal and non-verbal tasks and measures. Non-verbal measures include responses in tasks that do not involve any language input or output (categorization, memory of visual stimuli), accompanied by on-line measures, such as reaction times and eye-tracking, indicating to what particular features of events participants pay attention (eye-gaze) and how long they look at it (gaze duration) during the exploration of visual stimuli. ...
Previous research on motion expression indicates that typological properties influence how speakers select and express information in discourse (Slobin, 2004; Talmy, 2000). The present study further addresses this question by examining the expression of caused motion by adults and children (three to ten years) in French ( Verb-framed ) vs. English and German ( Satellite-framed ). Participants narrated short animated cartoons showing an agent displacing objects and varying along several dimensions (Path, Manner). A significant increase with age was found in the number of expressed motion components in all languages, as well as an influence of Path (vertical > boundary crossing). However, at all ages, participants encoded more information in English and German than in French, where more variation and structural changes occurred with increasing age. These findings highlight both cognitive and typological factors impacting the expression of caused motion in development. Implications of our findings are sketched in the ‘Discussion’.
This study explores the effects of instruction on the acquisition of motion event construal among learners of English as a second language. The challenge for learners with Verb-framed first languages is that they need to ‘unlearn’ the boundary-crossing constraint and conflate manner and motion in the main verb, as in she ran into the bank , however, there is little research on how this domain can be taught. We evaluate performance on story-telling productive tasks using three experimental treatments involving 1) an input-only approach based on the principles of Processing Instruction, 2) combined input and output training and 3) explicit information only about the target construction. The findings show that boundary-crossing constructions expressing manner can be taught and learning effects generalised to non-boundary-crossing structures not included in the training material. The effectiveness of input-only instruction persists over a two-week period, and compares positively with that of an input+output teaching package.
Full-text available
In recent years, considerable attention has been paid to languages that cannot be adequately described in Leonard Talmy's traditional framework of Satellite-framed and Verb-framed languages, resulting in cline-based and construction-based typologies. In the current paper, we focus on Greek, which has been said to have both Satellite-framed and Verb-framed characteristics. We compare two datasets, one experimental and one corpus-based, to uncover the impact of coding decisions and the implications for the classification of Greek as compared to Satellite-framed English and Verb-framed French. We situate the results from these two datasets in a wider analysis of motion theory, and show that taking into account semantic, syntactic and morphological aspects fares better than the exclusive focus on lexicalization patterns which was common in earlier work. We demonstrate the impact of the method of data type and the coding schemes on the characterization of linguistic patterns.
Full-text available
This paper examines whether cross-linguistic differences in motion encoding affect event processing, specifically memory performance. We compared speakers of two languages which differ strikingly in how they habitually encode MANNER and PATH of motion (Talmy in Toward a cognitive semantics: typology and process in concept structuring, 2nd edn, vol 2. MIT Press, Cambridge, 2000). We tested French and English adult native speakers across three tasks that recruited and/or suppressed verbal processing to different extents: verbal event descriptions elicited on the basis of dynamic motion stimuli, a verbal memory task testing the impact of prior verbalisation on target recognition, and a non-verbal memory task, using a dual-task paradigm to suppress internal verbalisation. Results showed significant group differences in the verbal description task, which mirrored expected typological tendencies. English speakers more frequently expressed both MANNER and PATH information than French speakers, who produced more descriptions encoding either PATH or MANNER alone. However, these differences in linguistic encoding did not significantly affect speakers' memory performance in the memory recognition tasks, neither in the verbal nor in the non-verbal condition. The findings contribute to current debates regarding the conditions under which language effects occur and the relative weight of language-specific and universal constraints on spatial cognition.
Full-text available
This study examines the impact of typological properties (satellite - vs. verb framed languages) on the expression of caused motion during adult second language acquisition. Productions were elicited by means of animated cartoons from 24 English learners of French (12 low-intermediate, 12 advanced) as compared to 24 native speakers (12 English, 12 French). The responses of native speakers differed with respect to semantic density (English>French) and to the systematic (English) vs. variable (French) devices used. As for learners, their utterance density increased with proficiency level as they acquired complex structures. Source/target language properties influenced this process, as shown by their increasing attempt to produce target-like structures that nonetheless remained source-like at both proficiency levels. These typological constraints suggest that learners do not construct an entirely independent linguistic system during second language acquisition and that L2 mastery may require some reconceptualization of spatial information. The discussion indicates research directions that might explore the implications of these results for language teaching. Keywords : second language acquisition, typology, space, caused motion
Languages vary strikingly in how they encode motion events. In some languages (e.g. English), manner of motion is typically encoded within the verb, while direction of motion information appears in modifiers. In other languages (e.g. Greek), the verb usually encodes the direction of motion, while the manner information is encoded in modifiers. We designed two studies to investigate whether these language-specific patterns affect speakers’ reasoning about motion. We compared the performance of English and Greek children and adults (a) in non-linguistic (memory and categorization) tasks involving motion events, and (b) in their linguistic descriptions of these same motion events. Even though the two linguistic groups differed significantly in terms of their linguistic preferences, their performance in the non-linguistic tasks was identical. More surprisingly, the linguistic descriptions given by subjects within language also failed to correlate consistently with their memory and categorization performance in the relevant regards. For the domain studied, these results are consistent with the view that conceptual development and organization are largely independent of language-specific labeling practices. The discussion emphasizes that the necessarily sketchy nature of speech assures that it will be at best a crude index of thought.
This volume offers theoretical and empirical contributions from varied disciplines addressing central questions at the forefront of the current literature in the domain of space: the nature of space in language, the variability of spatial systems across languages and its impact on language and cognitive development. It includes contributions from five fields of study, all of which are interested in closely related questions concerning spatial language and spatial cognition:- linguistics (including linguistic theory, typology, diachrony, sign-language);- cognitive/linguistic anthropology (cultural/linguistic relativity);- psycholinguistics (including language acquisition);- neuropsycholinguistics (language pathologies);- the philosophy of language.The following summary highlights some of the main general findings reported: (1)linguistic systems vary a great deal in how they represent space, but this variation can be organized around a relatively small set of types or parameters; (2)from the point of view of linguistic theory, space is not a primitive and homogeneous category; (3)data from language pathologies support the hypothesis that spatial cognition involves two different systems, one devoted to spatial language and the other to non-linguistic representations, and that these systems only partially overlap;(4)data from developmental psycholinguistics also show that the particular properties of linguistic systems have a strong impact on first language acquisition and suggest that they may also affect cognitive development.