The ‘thinking’ in thinking-for-speaking: Where is it?

Abstract

According to the thinking-for-speaking (TFS) hypothesis, speakers of different languages think differently while in the process of mentally preparing content for speech. The aim of the present paper is to critically discuss the research carried out within the TFS paradigm, against the background of the basic tenets laid out by the proponents of this framework. We will show that despite substantial progress in the investigation of crosslinguistic differences in the organisation of information in discourse, the studies that actually examine the cognitive aspects of speech production are, to date, vanishingly few. This state of affairs creates a gap in our knowledge about the thought processes that co-occur with speech production during language use and acquisition. We will argue that in order to reach a more comprehensive picture of the cognitive processes and outcomes of speech production, methodologies additional to the analysis of information organisation must be used.

Full text

is is a contribution from Language, Interaction and Acquisition 4:1
© 2013. John Benjamins Publishing Company
is electronic le may not be altered in any way.
e author(s) of this article is/are permitted to use this PDF le to generate printed copies to
be used by way of oprints, for their personal use only.
Permission is granted by the publishers to post this le on a closed server which is accessible
to members (students and sta) only of the author’s/s’ institute, it is not permitted to post
this PDF on the open internet.
For any other use of this material prior written permission should be obtained from the
publishers or through the Copyright Clearance Center (for USA: www.copyright.com).
Please contact rights@benjamins.nl or consult our website: www.benjamins.com
Tables of Contents, abstracts and guidelines are available at www.benjamins.com
John Benjamins Publishing Company

Language, Interaction and Acquisition : (), 91–100.  ./lia...ath
 – / - – © John Benjamins Publishing Company
e ‘thinking’ in thinking-for-speaking
Where is it?
Panos Athanasopoulos and Emanuel Bylund
Reading University / Stockholm University
According to the thinking-for-speaking (TFS) hypothesis, speakers of dierent
languages think dierently while in the process of mentally preparing content
for speech. e aim of the present paper is to critically discuss the research car-
ried out within the TFS paradigm, against the background of the basic tenets laid
out by the proponents of this framework. We will show that despite substantial
progress in the investigation of crosslinguistic dierences in the organisation of
information in discourse, the studies that actually examine the cognitive aspects
of speech production are, to date, vanishingly few. is state of aairs creates a
gap in our knowledge about the thought processes that co-occur with speech
production during language use and acquisition. We will argue that in order to
reach a more comprehensive picture of the cognitive processes and outcomes
of speech production, methodologies additional to the analysis of information
organisation must be used.
Keywords: inking-for-speaking, conceptualisation, linguistic relativity
1. Introduction
e idea that speakers of dierent languages think dierently has been subject to
scholarly debate and empirical inquiry in a range of dierent disciplines. ough
this question was discussed already in the ancient schools of Greek and Sanskrit
philosophy, it is commonly associated with the linguist Benjamin Lee Whorf, who
formulated the so-called linguistic relativity principle. According to this principle,
speakers of dierent languages are not equal observers of the world, but are point-
ed by their grammars towards dierent evaluations of reality (Whorf 1956: 241).
Research into linguistic relativity thus seeks to establish whether crosslinguistic
dierences in the semantic partitioning of reality give rise to crosslinguistic dier-
ences in thought. A cornerstone in this line of inquiry is the denition of ‘thought’

© 2013. John Benjamins Publishing Company
All rights reserved
92 Panos Athanasopoulos and Emanuel Bylund
as non-verbal behaviour, operationalised typically as a range of dierent cognitive
processes, such as reasoning, classication, and categorical perception (Lucy 1992,
1997).
e linguistic relativity principle is, however, not the only account that pro-
vides a framework for research into the relationship between language and
thought. In more recent years, Dan Slobin (1987, 1991, 1996, 2003) formulated
the thinking-for-speaking (henceforth TFS) hypothesis as an alternative to the lin-
guistic relativity principle. According to the TFS hypothesis, speakers of dierent
languages think dierently while in the process of mentally preparing content for
speech. e crucial dierence between LR and TFS is that the former focuses on
eects of linguistic structure on non-verbal behaviour and conceptual representa-
tion, while the latter focuses on eects of linguistic structure on the cognitive pro-
cesses involved in speech production, i.e. the conceptualisation stage in Levelt’s
(1989) terms. More specically, Slobin suggests that the speaker attends to and
verbalises those aspects of reality that are readily encodable in his/her language.
e past decade has seen a steady increase in studies conducted within the TFS
framework, and there is a growing number of investigations examining TFS in
dierent types of bilingual speakers (for recent overviews and contributions, see
Benazzo, Flecken & Soroli 2012; Cook & Bassetti 2011; Han & Cadierno 2011;
Jarvis 2011; Pavlenko 2011).
e aim of the current paper is to critically discuss the research carried out
within the TFS paradigm, against the background of the basic tenets laid out by
the proponents of this framework. We will show that whereas there is a large num-
ber of crosslinguistic studies on the organisation of information in discourse, the
studies that actually examine the cognitive aspects of speech production are, to
date, vanishingly few. is bias creates a gap in our knowledge about the thought
processes that co-occur with speech production, and reduces the ‘thinking’ com-
ponent in TFS to a mere question of information structure. We will argue that in
order to reach a more comprehensive picture of the cognitive processes and out-
comes of speech production, methodologies additional to the analysis of informa-
tion organisation must be used.
2. e thinking-for-speaking framework
e TFS framework postulates that the activity of thinking takes on a particular
quality when used in the activity of speaking (Slobin 1996: 76). is cognitive pro-
cess of thinking for the purpose of speaking constitutes, in other words, a “special
form of thought that is mobilised for communication” (ibid.). In essence, when
thinking for speaking, the speaker chooses those characteristics of events and

© 2013. John Benjamins Publishing Company
All rights reserved
e ‘thinking’ in thinking-for-speaking 93
objects that, rst, t some conceptualisation of the event, and second, are read-
ily encodable in his/her language (ibid.). In concrete terms, this means that, for
example, when talking about a given event, a speaker of Turkish will have to check
the properties of the information source underlying the predication, evidentiality
being a grammatical category in Turkish, or that a speaker of Russian will have
to check the temporal distribution of the event, since aspect is grammaticised in
Russian.
e rationale underlying the TFS framework is that human beings are most
of the time engaged in preparing, producing, or interpreting verbal messages, and
therefore research into language and thought is incomplete without attention to
the thought processes that relate to speech production. It is suggested that TFS
eects can be documented in dierent time frames in relation to speech produc-
tion (Slobin 2003: 183 et passim): the rst frame concerns prelinguistic or non-
linguistic coding, which occurs when the speaker attends to those dimensions of
reality that are relevant for subsequent linguistic coding. ese eects are called
‘anticipatory’ eects. e second time frame is the speaking time proper, which
is when the linguistically codable dimensions are transformed into speech (this
time frame also applies to speech comprehension). e third frame is situated past
speech production or comprehension, and relates to eects of the linguistic coding
of reality on subsequent recall or recognition. ese eects are called ‘consequen-
tial’ eects.
Slobin (1987) suggests that TFS can be divided into several subtopics, one of
which concerns rst language (L1) acquisition. For the past two decades, research-
ers have collected data on information structure in child and adult speakers in a
number of dierent languages, with the intention to establish the emergence of
language-specic patterns of TFS in the course of L1 acquisition (e.g., Berman &
Slobin 1994; Strömqvist & Verhoeven 2004). Another TFS topic that is increas-
ingly receiving attention relates to the phenomena of additional language learn-
ing. Here, the question at issue concerns whether or to what extent an additional
language learner can attain the TFS patterns of the target language (e.g., Benazzo
et al. 2012; Verhoeven & Strömqvist 2001).
3. e current state of TFS: Too much emphasis on the speaking,
little attention to the thinking
What kinds of questions might be asked in a research agenda linking language
and thought in native speakers and in bilinguals/L2 users? One aim would be to
investigate how people think about and construct conceptual representations of
objects, events, colours, spatial relations, time and the like as a function of their

© 2013. John Benjamins Publishing Company
All rights reserved
94 Panos Athanasopoulos and Emanuel Bylund
native/second language. A second aim would be to address the thought processes
taking place during speech production, probing the way speakers develop strat-
egies of encoding and organising information during online speech production
(or comprehension). e rst aim is in fact at the heart of modern enquiries into
the linguistic relativity hypothesis. As such, to address this aim, studies typically
employ measures of nonverbal performance (e.g. categorisation, similarity judge-
ments, recognition memory, scene comprehension). e second aim is closer to
Slobin’s TFS. According to Slobin (2003), to fully address this aim, some sort of
methodology that would elucidate the temporal dynamics of developing strategies
during speaking should be employed. Such methodologies might include speech
onset time and comprehension reaction times, or some behaviour concurrent to
speech production and/or comprehension (e.g. visual attention allocation, gesture,
event-related brain potentials).
e rst point we wish to raise in relation to the above aims is the importance
of treating linguistic relativity and TFS as distinct (albeit conceptually related) hy-
potheses. While the cornerstone of both hypotheses is the fact of linguistic diversity
in semantics and grammar, the methodologies employed to address each question
are necessarily dierent (see previous paragraph and also Lucy 1997; Odlin 2011;
Pavlenko 2011; Treers-Daller 2012). From a theoretical viewpoint, it is important
to distinguish these two hypotheses because they concern two dierent levels of
mental representation. As Lucy (1997) and others have argued, phenomena re-
lated to selecting and structuring content for speech as captured by the thinking-
for-speaking paradigm concern a linguistic level of representation distinct from a
non-linguistic conceptual level that concerns cognitive representation of concepts,
which may be language-derived to a lesser or greater extent depending on the
conceptual category, the task, etc. (cf. Athanasopoulos & Bylund, 2013). Any kind
of linguistic inuence on that level of representation may be taken as evidence
for linguistic relativity (Lucy 1992, 1997). ere is ample empirical evidence to
support the distinction between these two levels of processing. For example, stud-
ies show that cognitive representation of motion and thinking-for-speaking about
motion are largely independent from each other. Across many dierent experi-
ments, apparent cross-linguistic dierences are found when talking about events,
but not when cognitively processing motion scenes outside of overt verbalisation
(Gennari, Sloman, Malt & Fitch 2002; Papafragou, Massey & Gleitman 2002) or
memory-based cognitive processing (Papafragou & Selimis 2010; Trueswell &
Papafragou 2010; Athanasopoulos & Bylund, 2013; Bylund, Athanasopoulos &
Ooostendorp, in press). Clearly, the degree to which thinking-for-speaking de-
pends on language-specic grammatical categories is orthogonal to the degree to
which conceptual representations themselves are language-independent or not.

© 2013. John Benjamins Publishing Company
All rights reserved
e ‘thinking’ in thinking-for-speaking 95
e second point we wish to put across in relation to the aims of this section,
and most central for the purposes of the entire paper, is the observation that many
studies purporting to investigate TFS per se do not in fact provide sucient data
on thought processes during speaking, but only describe linguistic diversity in
the sense that they report typologically-constrained verbalisations produced by
speakers of dierent languages. is, then, is taken as evidence of crosslinguistic
dierences in the thought processes concurrent to speech production. It is unclear,
however, how much can be inferred about co-verbal thought processes from pure-
ly verbal data. In fact, this problem of inference resembles that which many early
approaches to linguistic relativity faced, in which linguistic diversity was not only
the premise, but also the nal evidence of crosslinguistic dierences in cognition.
Such circular reasoning is parodied by Pinker (1994: 61) as follows: “[Eskimos]
speak dierently so they must think dierently. How do we know that they think
dierently? Just listen to how they speak!” Applying this logic to research into TFS,
one could ask what evidence there is to assume that speakers of dierent languages
think dierently while describing a given scene because of the mere fact that they
describe the scene dierently. If the only evidence for this assumption is the cross-
linguistic dierences in information structure about the scene, then there is a clear
component of circularity to this reasoning.
e literature abounds of studies examining semantic and syntactic aspects of
information organisation from a crosslinguistic perspective (see contributions in,
e.g., Benazzo et al. 2012; Berman & Slobin 1994; Han & Cadierno 2011; Pavlenko
2011; Strömqvist & Verhoeven 2004). e typical approach of studies of this kind
is to compare verbal descriptions of stimuli amongst native speakers of dierent
languages and/or language learners. e typical nding is that people tend to say
what is typologically possible (in the case of obligatory linguistic constraints) and/
or prototypical (in the case of non-obligatory but preferred linguistic features) in
their native language, while L2 learners may show pervasive inuence of the L1 in
their L2 productions, but also in some cases inuence of the L2 on the L1. While
these studies may provide sophisticated linguistic analyses of information focus,
semantic density, semantic locus, syntactic architecture and the such, they rely on
speech elicitation methods that reveal little or nothing about the online psycho-
linguistic processes and dynamic conceptualisation strategies at play. erefore,
claims made about crosslinguistic dierences in conceptual representation remain
speculative. Production data alone reveal what people have said, and as such ad-
dress only one aspect of Slobin’s thinking-for-speaking hypothesis (speaking time
proper). Such data tell us very little about how and why speakers choose to encode
information in a particular way (Slobin’s anticipatory stage), and little focus is
paid to consequential eects on cognition of immediately preceding verbalisa-
tion (Slobin’s third stage of thinking-for-speaking, see Gennari et al. 2002 for an

© 2013. John Benjamins Publishing Company
All rights reserved
96 Panos Athanasopoulos and Emanuel Bylund
empirical approach). Because of the dynamic nature of speech production, the use
of online methods of speech production performance (see next section for exam-
ples) would allow for a more complete picture of thinking-for-speaking processes.
4. Probing the thinking in TFS: Towards a new research agenda
In this section, we will argue for a methodological shi in focus from surface ver-
bal production to co-verbal behaviours such as gestures, attention allocation dur-
ing speech, and ERPs in semantic comprehension, because we believe it is in these
co-verbal behaviours that the ‘thinking’ in thinking-for-speaking may be most
usefully conceptualised, and empirically operationalised. e caveat of inferring
information about conceptualisation processes from speech data alone is high-
lighted by the increasing amount of studies that have emerged in the past ve years
or so, showing that when dynamic measures of online behaviour are used, those
measures can oen reveal conceptualisation processes that are at variance with, or
nuance further, surface production patterns.
For example, Flecken (2011) investigated eects of grammatical aspect on
verbal encoding in early bilingual speakers of Dutch (an aspect language) and
German (a non-aspect language). Participants had to describe a set of video clips
focusing on the question ‘What is happening?’ Additionally, their eye movements
were recorded in order to study their planning and organisation of content that
was going to be expressed. e results showed that bilinguals displayed patterns of
language use that were in-between the two monolingual norms. Specically, these
early bilinguals frequently used the progressive form in Dutch, but they also tend-
ed to combine progressive aspect with the mention of endpoints, a combination
that is not at all frequent in monolingual Dutch speakers. e eye-tracking analy-
sis revealed that the Dutch–German bilinguals, when speaking Dutch, have an
underlying bilingual-specic event conceptualisation system that diers not only
from that of Dutch monolinguals, but that also shows no sign of cross-linguistic
inuence from the monolingual German pattern.
Brown and Gullberg (2008) report dissociations in surface production and
gesture behaviour in Japanese and English monolinguals and L2 users. When
Japanese speakers mention manner of motion in speech, they are most likely to
also gesture about manner. However, when English speakers mention manner in
speech, they oen gesture only about the path of motion. Japanese L2 English
users were signicantly more likely to talk about manner and gesture about path
in their L1 Japanese than were monolingual Japanese speakers, thus display-
ing English-like conceptualization in their L1 productions. More strikingly, in a
study of placement descriptions in the speech and gestures of Dutch and German

© 2013. John Benjamins Publishing Company
All rights reserved
e ‘thinking’ in thinking-for-speaking 97
learners of L2 French, Gullberg (2011) found that both L2 groups produced target-
like spoken L2 French with respect to the linguistic categories under scrutiny, sug-
gesting restructuring of conceptualisation patterns as a function of L2 acquisition.
However, when Gullberg (2011) examined the accompanying gesture behaviour
of the learners, she found that both L2 groups showed evidence of L1 inuence on
conceptualisation, with many learners showing an in-between pattern, displaying
both L1 and L2 gesture patterns within the same utterance.
To provide an example from comprehension processes, in a study looking at
the eects of grammatical gender while having to make decisions about the se-
mantic relationships between objects, Boutonnet, Athanasopoulos and ierry
(2012) tested Spanish-English bilinguals and native speakers of English in a se-
mantic categorisation task. Participants were presented with triplets of pictures in
an all-in-English context while measuring event-related brain potentials (ERPs).
Participants were asked to press one button when the third picture of a triplet
belonged to the same semantic category as the rst two, and another button when
it belonged to a dierent category. Unbeknownst to them, in half of the trials, the
gender of the third picture name in Spanish had the same gender as that of the
rst two, and the opposite gender in the other half. Behavioural results showed no
measurable eect of gender consistency in either of the participant groups. at
is, Spanish-English bilinguals paid no more attention to gender incongruences in
the triplets of stimuli than English monolinguals did. In contrast, ERPs revealed
not only the expected semantic priming eect in both groups, but also a dier-
ent brainwave pattern, modulated by gender inconsistency in Spanish-English bi-
linguals, exclusively. ese results show that these bilinguals spontaneously and
unconsciously access grammatical gender in their L1, even though on the surface
they perform like native speakers of their L2.
In a study combining production patterns, attention allocation, and compre-
hension of scenes from memory, von Stutterheim, Andermann, Carroll, Flecken,
and Schmiedtova (2012) showed that speakers of [+aspect] languages (e.g. Arabic,
Spanish, English) who do not typically mention the endpoint of an action (Von
Stutterheim & Nüse, 2003), also tend to look less at endpoints, and do not store
information on endpoints to the same extent as speakers of [−aspect] languages
(e.g. Czech, Dutch, German) do. e nding that speakers of [+aspect] languages
mention more endpoints than speakers of [−aspect] languages shows an eect of
linguistic typology on what speakers choose to mention. What shed light on the
actual process of speech planning was the eye-tracking analysis: since the end-
point will not be relevant in planning what to say, speakers of [+ aspect] languages
directed less visual attention to the endpoint-object in the stimulus than speakers
of [−aspect] languages. Eye-tracking analyses also revealed that the eect of lin-
guistic typology on conceptualisation is not an all-or-nothing phenomenon. Even

© 2013. John Benjamins Publishing Company
All rights reserved
98 Panos Athanasopoulos and Emanuel Bylund
though speakers of [+ aspect] languages mentioned fewer endpoints than speak-
ers of [−aspect] languages (as predicted by linguistic typology), visual attention in
speakers of [+ aspect] languages was directed to the potential endpoint to some
extent. e crucial dierence between populations seems to be that [+ aspect]
speakers look at the endpoints at a later point, compared to [−aspect] speakers (cf.
von Stutterheim & Carroll 2006). In other words, it is necessary to also take into
account the time course of visual attention as an additional window on conceptual
processing. While [−aspect] speakers may direct attention to the endpoint region
before speech onset, [+aspect] speakers may do so in the course of utterance ar-
ticulation (von Stutterheim et al. 2012).
5. Conclusions
Linguistic diversity is the cornerstone of both the linguistic relativity hypothesis
and the thinking-for-speaking hypothesis. However, linguistic diversity can never,
in and of itself, be taken at face value as evidence for diversity in thinking (Lucy
1997) or in thinking-for-speaking (Slobin 2003). Purely linguistic analyses are a
fundamental starting point for TFS, but at the same time a tad rudimentary as an
only tool for ferreting out the cognitive processes involved in speech production.
We have argued that the vast majority of studies look at the organisation of infor-
mation structure in discourse when addressing TFS. But information structure is
the end result of a complex process. Studies tend to overlook this process, or make
assumptions about this process by looking at the end result, rather than at the pro-
cess itself. Looking at speech data alone can tell us what people choose to say, but
does not reveal how they come to say it. In other words, they leave the big question
of the cognitive processes involved in preparing content for speech unanswered.
Acknowledgements
e work with this paper was supported by the Swedish Research Council (grant nr 421-2010-
2104).
References
Athanasopoulos, P. & Bylund, E. (2013). Does grammatical aspect aect motion event cogni-
tion? A cross-linguistic comparison of English and Swedish speakers. Cognitive Science, 37,
286–309.

© 2013. John Benjamins Publishing Company
All rights reserved
e ‘thinking’ in thinking-for-speaking 99
Benazzo, S., Flecken, M. & Soroli, E. (eds.) (2012). Typological perspectives on second lan-
guage acquisition: ‘inking for Speaking’ in L2. Special issue of Language, Interaction, and
Acquisition, 3.
Berman, R. & Slobin, D. (eds.) (1994). Relating events in narrative. Hillsdale, NJ: Lawrence
Erlbaum.
Boutonnet, B., Athanasopoulos, P. & ierry, G. (2012). Unconscious eects of grammatical
gender during object categorisation. Brain Research 1479, 72–79.
Brown, A. & Gullberg, M. (2008). Bidirectional crosslinguistic inuence in L1-L2 encoding of
manner in speech and gesture: A study of Japanese speakers of English. Studies in Second
Language Acquisition 30, 225–251.
Bylund, E., Athanasopoulos, P. & Ooostendorp, M. (in press). Motion event cognition and
grammatical aspect: Evidence from Afrikaans. Linguistics.
Cook, V. & Bassetti, B. (eds.) (2011). Language and bilingual cognition. Hove, UK: Psychology
Press.
Flecken, M. (2011). Event conceptualization by early Dutch-German bilinguals: Insights from
linguistic and eye-tracking data. Bilingualism: Language and Cognition 14, 61–77.
Gennari, S., Sloman, S., Malt, B. & Fitch, W. (2002). Motion events in language and cognition.
Cognition 83, 49–79.
Gullberg, M. (2011). inking, speaking, and gesturing about motion in more than one lan-
guage. In A. Pavlenko (ed.), inking and speaking in two languages (143–169). Bristol:
Multilingual Matters.
Han, Z. & Cadierno, T. (eds.) (2011). Linguistic relativity in L2 acquisition: thinking for speaking.
Bristol: Multilingual Matters.
Jarvis, S. (ed.) (2011). Crosslinguistic inuence in bilinguals’ concepts and conceptualizations.
Special issue of Bilingualism: Language and Cognition, 14.
Levelt, W. (1989). Speaking: from intention to articulation. Massachusetts: MIT Press.
Lucy, J.A. (1992). Grammatical categories and cognition: a case study of the linguistic relativity
hypothesis. Cambridge: Cambridge University Press.
Lucy, J.A. (1997). Linguistic relativity. Annual Review of Anthropology 26, 291–312.
Odlin, T. (2011). On the interdependence of conceptual transfer and relativity studies. In Z. Han
& T. Cadierno (eds.), Linguistic relativity in L2 acquisition: inking for speaking (183–194).
Bristol: Multilingual Matters.
Papafragou, A., Massey C. & Gleitman, L. (2002). Shake, rattle, ‘n’ roll: e representation of
motion in language and cognition. Cognition 84, 189–219.
Papafragou, A. & Selimis, S. (2010). Event categorisation and language: A cross-linguistic study
of motion. Language and Cognitive Processes 25, 224–260.
Pavlenko, A. (ed.) (2011). inking and speaking in two languages. Clevedon: Multilingual
Matters.
Pinker, S. (1994). e language instinct. New York: William Morrow.Slobin, D. (1987). inking
for speaking. Proceedings of the Annual Meeting of the Berkeley Linguistics Society 13, 435–
444.
Slobin, D. (1991). Learning to think for speaking: Native language, cognition, and rhetorical
style. Pragmatics 1(1), 7–25.
Slobin, D. (1996). From “thought and language” to “thinking for speaking.” In J. Gumperz & S.
Levinson (eds.), Rethinking linguistic relativity (70–96). Cambridge: Cambridge University
Press.

© 2013. John Benjamins Publishing Company
All rights reserved
100 Panos Athanasopoulos and Emanuel Bylund
Slobin, D. (2003). Language and thought online: cognitive consequences of linguistic relativ-
ity. In D. Gentner & S. Goldin-Meadow (eds.), Language in mind: Advances in the study of
language and thought (157–192). Cambridge, MA: MIT Press.
Strömqvist, S. & Verhoeven, L. (eds.) (2004). Relating events in narrative. Vol II. Typological and
contextual perspectives. Mahwah, NJ: Lawrence Erlbaum.
von Stutterheim, C., Andermann, C., Carroll, M., Flecken, M. & Schmiedtova, B. (2012). How
grammaticized concepts shape event conceptualization in language production: Insights
from linguistic analyses, eye tracking data, and memory performance. Linguistics 50, 833–
867.
von Stutterheim, C. & Carroll, M. (2006). e impact of grammaticalised temporal categories
on ultimate attainment in advanced L2-acquisition. In H. Byrnes, H.D. Weger-Gunthrap &
K. Sprang (eds.), Educating for advanced foreign language capacities: Constructs, curriculum,
instruction, assessment (40–53). Washington, DC: Georgetown University Press.
von Stutterheim, C. & Nüse, R. (2003). Processes of conceptualization in language production:
Language-specic perspectives and event construal. Linguistics 41, 851–881.
Treers-Daller, J. (2012). inking for speaking and linguistic relativity among bilinguals.
Language, Interaction, and Acquisition 3, 288–300.
Trueswell, J.C. & Papafragou, A. (2010). Perceiving and remembering events cross linguistically:
Evidence from dual-task paradigms. Journal of Memory and Language 63, 64–82.
Verhoeven, L. & Strömqvist, S. (eds.) (2001). Narrative development in a multilingual context.
Amsterdam: John Benjamins.
Whorf, B.L. (1956). Language, thought, and reality. Selected writings of Benjamin Lee Whorf (J.B.
Carroll, ed.). Cambridge, MA: MIT Press.
Résumé
Selon l’ hypothèse du “Penser-pour-parler” (inking-for-speaking ou TFS), les locuteurs de
langues diérentes n’organiseraient pas leur pensée de la même manière lors de la préparation
mentale de leurs productions. L’objectif de cet article est de proposer une discussion critique
des travaux eectués dans le paradigme du TFS, à partir des principes de base établis par ce
cadre théorique. Nous montrons que, malgré les avancées importantes qu’a apportées l’étude
des diérences inter-langues dans l’organisation du discours, les travaux examinant réellement
les aspects cognitifs de la production de la parole sont encore très rares. A partir de ce bilan,
nous montrons l’insusance de nos connaissances concernant les processus cognitifs qui ac-
compagnent la production pendant l’utilisation et l’acquisition du langage. Hormis l’analyse
de l’organisation discursive, nous concluons que d’autres méthodologies sont nécessaires an
d’obtenir une vision plus globale des processus cognitifs présidant à la production du langage et
des implications qui en émergent.