Rafał Szymanek
University of Amsterdam
MSc in Brain and Cognitive Sciences: cognitive neuroscience track
How speaking two languages affects
cognition:
A review of empirical evidence for linguistic relativity in bilingual studies
ABSTRACT
The central claim of the linguistic relativity hypothesis (Whorf, 1940) is that the grammatical and semantic characteristics of language affect how we perceive and think about the external world. This literature review examines what happens when a person speaks two languages with conflicting characteristics. Bilingual studies from several linguistic domains are discussed in order to investigate changes in cognition and mental conceptualization of the world related to multiple language knowledge and use. The empirical data is framed in the context of the conceptual framework presented by Gentner and Goldin-Meadow (2003), which describes the effects of language on cognition in terms of whether they affect perception, categorization, or conceptualisation of the external world. The studies are further compared with different scenarios of mental concept formation described by Bassetti and Cook (2011). The picture of the bilingual mind presented here addresses several inconsistencies and challenges present in the current approaches and underlines the importance of conducting bilingual research in the field of language cognition.
Student ID: 11119438
January 2017
Supervisor: prof. Annette de Groot
Co-assessor: prof. Jeannette Schaeffer
Introduction
Linguistic diversity and linguistic relativity
At present, the over 7 billion people living on Earth communicate with each other using one of the around 7000 different languages (Lewis, Simons, & Fennig, 2009), none of which is mutually intelligible with more than a handful of others. The diversity of languages has been of interest to linguists, who look at cross-language syntactical and semantic differences, and to psychologists, who try to reconstruct the mental concept space of speakers of different languages and the processes leading to language comprehension and production.
Languages differ on many levels such as in the speech sounds they use, the script they are written with, or how they arrange words within a sentence. For the purpose of this paper, I will leave some of these differences out, focussing on grammatical aspects of language that have semantic consequences, such as the use of grammatical gender and number when dealing with nouns as well as the use of grammatical tense and aspect in verbs. Furthermore, I will look at how languages express motion and spatial relations, and how they classify objects into categories, using examples from phenomena such as colour perception and object classification.
It has been hypothesised that differences between languages cause differences in how their speakers think, an idea known as linguistic relativity. Sapir and Whorf were the first to bring this topic into mainstream psychology research (Whorf, 1940), giving it its other name: the Sapir-Whorf hypothesis. In its extreme version, termed linguistic determinism, language determines the content of thought and defines the limits of perception. Even though the extreme version of the hypothesis has been frequently challenged (e.g. Devitt & Sterelny, 1999; Pinker, 1994), there are now countless examples of experimental results which are in line with the notion contained in its weaker version, linguistic relativity.
Language can shape cognition in myriad ways. The way that actions are encoded in language directs how speakers think about them: whether a language uses grammatical aspect to distinguish between completed and uncompleted actions affects how speakers remember and describe events. It has been found (Bylund, Athanasopoulos, & Oostendorp, 2013) that when describing motion events, such as a scene where a man is walking in the direction of a house situated in the background, speakers of Afrikaans and Swedish (non-aspect languages) pay more attention to the end-point (house) than do speakers of English (an aspect language). This is because speakers of aspect languages, due to their use of the perfective aspect, are attuned to incomplete, on-going actions and might encode the scene as simply “a man is walking” instead of “a man walks towards a house”. Similarly, Spanish speakers are more attuned to the path of motion (e.g. in vs. out) whereas English speakers to its manner (e.g. walk vs. run) because of the specific feature that is more salient in motion expressions in their respective language (Slobin, 1996).
How objects are encoded in language also has an influence on how they are perceived. In some languages, such as Spanish and German, grammatical gender is obligatory for all nouns. However, because grammatical gender is an arbitrary feature of object nouns, the same object can have a masculine gender in one language but a feminine one in the other (Foundalis, 2002). Speakers follow their language’s gender assignments when asked to assign a sex to an object or to describe it, and when assigning a male or female
voice to a talking object (Sera, Berge, & Pintado, 1994). So a bridge will be described using adjectives with feminine connotations (such as elegant) by Germans and using masculine adjectives (such as strong) by Spaniards, because of the gender assignment to that particular noun in their respective languages (Bassetti, 2007). Similarly, Lucy (1992) has shown that Yucatec speakers, whose language does not assign grammatical number to non-animate objects such as bottles, consider two pictures differing only in the number of bottles presented as more similar than do English speakers. Because their language does not treat bottles as discrete entities, they are not giving much significance to those differences.
Lens, toolkit, and category maker
The ways in which language affects cognition have been classified by Gentner and Goldin-Meadow (2003) into three types: language as lens, toolkit, and category-maker.
Language as lens refers to the effect of language on the perception of the world. This includes most of the examples discussed in the previous section, such as ascribing masculine characteristics to an object because of the masculine gender assignment for that particular object in a given language or paying more attention to the path or manner of a movement depending on which of these aspects is more transparent in a given language. This type of effects is the closest to the original Sapir-Whorf hypothesis and will be, in fact, the main focus of the remainder of this paper.
Language as toolkit refers to how the acquired language influences the capacity for representation and reasoning, in other words, to how being a speaker of a language affects your mental processes. This aspect of linguistic relativity relates more to the general effect of speaking a language than to differences between speakers of specific languages and is thus both less interesting for our present purposes and more difficult to study empirically. It is worth noting, however, that this approach leads to some interesting empirical avenues, such as testing a pair of languages, of which one possesses a specific grammatical feature, that could be used as a tool in dealing with daily situations, and the other does not, like the ergative case which forces speakers to think whether an actor is performing an action on its own or towards another agent or object (i.e. distinguish between transitive and intransitive verbs).
Finally, language as a category maker refers to how we cut the world into distinct representational categories based on the labels provided by our language. This type of influence overlaps, in fact, with language as lens effects, as it also relates to the way the perception of the world is affected by the language we speak. Some examples that I will look at in later sections include colour perception and object grouping.
These three broad and partly overlapping categories of linguistic effects on cognition can also be viewed as approaches towards linguistic relativity and can influence the methodology and interpretation of empirical studies. Because of this, they are useful to keep in mind when discussing experimental results, which I will do in the following sections.
Bilingualism
In order to fully grasp the extent of linguistic relativity empirically, other factors influencing cognition have to be controlled for. These factors include cultural and geographical influences that naturally go hand in hand with language. Consider the hunter-gatherer Hadza people speaking the Hadza language (Marlowe,
2010). They will think differently than a businessman from North America speaking English, but it is anyone’s guess how much of this difference stems from language alone. Therefore, research must focus on populations which minimize these contaminating factors. For this reason, the ideal populations to study linguistic relativity are bilinguals and multilinguals, that is, people who master more than a single language. For the purpose of this paper, I will define bilingualism broadly and consider any person able to speak more than one language a bilingual, whether they have spoken the second language from birth (as is the case with simultaneous bilinguals) or acquired it later in life (as in sequential bilinguals). I will also refer to second-language learners using this term, as research using this population has led to results similar to that of bilingual speakers in some domains, as I shall discuss further. Any time a bilingual of a specific language pair will be mentioned, I will put the languages in the order of acquisition, so a native French speaker who has learned Spanish later in life will be referred to as a “French-Spanish bilingual”. Effects of bilingualism on cognition can be divided into macro-effects, i.e. generic effects of bilingualism, and micro-effects, i.e. consequences of knowing specific languages. There are various macro-effects associated with bilingualism. Knowing two languages causes a widening of the speaker’s conceptual world, which in turn might lead to enhanced creativity (Ricciardelli, 1992). Furthermore, the fact that a person constantly uses two different languages means they always have to select between one of two possible means of expressing a concept. This leads to increased attentional control, causing bilinguals to perform better at solving problems that require inhibition of misleading information. Because of that, bilinguals are better at task switching and develop a theory of mind earlier than monolinguals (Bialystok, 2005). Bilingualism is also believed to improve cognitive functioning and aid divergent thinking (de Groot, 2011).
In the reminder of this paper I will present examples of micro-effects of bilingualism on cognition, focusing on speakers of specific language pairs. I will review experimental evidence for different performance on various cognitive tasks depending on the language used, and discuss what this can tell us about the functioning of the bilingual mind.
Empirical evidence for linguistic relativity in bilinguals
How grammar affects perception and cognition in bilinguals
I will first discuss influences arising from structural differences, due to the presence of various grammatical features in one language but not in the other, or implemented differently in different languages. I will show them to affect, for the most part, the lens of language, that is, how we perceive the world. Then, I will discuss semantic differences, coming about because of attributing different meaning to objects or because of using labels that apply to a different range of objects in each of the two languages. These differences mostly affect, as I will show in the next section, the category-maker aspect of language, that is, how we divide the stimuli perceived in the world into meaningful concepts.
Grammatical gender
In the previous sections I have shown how language shapes perception, adjusting the lens through which we look at reality. It follows that learning a second language should change the way we perceive the world. Indeed, Kurinski and Sera (2011) found that even as little as ten weeks of studying Spanish—a language
with obligatory grammatical gender markings—was enough to change the way that English speakers attribute gender to inanimate objects. They followed the progress of university students beginning a course in elementary Spanish and administered a test which consisted of assigning a male or female voice to a talking object. The experiment was carried out at various stages during the language course. They found that the more time the students spent on learning the second language, the more their assignments resembled those of native Spanish speakers, suggesting that participants have acquired the concept of grammatical gender and applied it to non-living objects.
This effect is even more pronounced in fully developed bilinguals. Phillips and Boroditsky (2003) investigated gender assignment in German-English and Spanish-English bilinguals in an object naming task. They asked participants to remember a proper name assigned to an artefact (e.g. an apple). The names were loaded with gender (e.g. Patrick or Patricia). The artefacts and names were assigned in such a way that some of the names agreed in terms of gender with the artefact’s grammatical gender in the participant’s language and some did not. They found that bilingual speakers were still affected by their first language’s gender assessment and thus performed better when the name was consistent with the gender of the item in their first language. Similarly, Scheutz and Eberhard (2004) showed that German-English bilinguals tend to associate male connotations to the German-English morpheme “-er”, which denotes masculinity in German but not in English, despite having a similar form and function. English speakers did not exhibit this behaviour, showing the influence of first-language associations on second-language sentence encoding.
However, the empirical evidence on gender is not entirely clear-cut. It is possible that in the studies mentioned above, participants were using grammatical gender as a conscious strategy. Since the tasks— which involved comparing inanimate objects to humans and assigning a human name to objects—were novel and abstract, it is possible that participants decided to use grammatical gender to solve an otherwise impossible problem (Kousta, Vinson, & Vigliocco, 2008). Furthermore, the aforementioned effects of first-language grammatical gender on attributing a male or female voice to an object were found to only occur in languages where the grammatical gender corresponds with natural gender, that is, languages that distinguish between masculine and feminine nouns (like Spanish). In languages that also possess a neuter class (like German), this effect has not been found (Sera, Elieff, Forbes, & Burch, 2002). This result supports the possibility that it is in fact a conscious strategy that is responsible for positive results and, since it is not possible to use that strategy with neuter nouns to assign a neuter-sounding voice to an object, it fails in the case of languages such as German (Kousta et al., 2008). Future research should control for that possibility, for example by including a comprehensive assessment of the strategies used by participants when performing these novel tasks.
Grammatical number
Language also determines how we dissect reality into different categories. One aspect of language that illustrates this phenomenon is grammatical number. English always assigns plural markings for animate objects (e.g. two cows) but inanimate objects can be either discrete, such as tools (two hammers), or not, such as substances (milk). Substances do not take plural markings but require an additional quantifier to express plurality (e.g. two glasses of milk but not *two milks). In other languages such as Yucatec (a Mexican language) or Japanese, animate objects always take plural markings and inanimate never do.
Monolingual speakers, as I showed in the previous sections, are affected by this: Whereas English speakers consider a difference in the number of tools more significant than a difference in the number of substances, Yucatec speakers do not (Lucy, 1992). More recently, Athanasopoulos (2006) investigated this phenomenon in a bilingual population of Japanese speakers with English as their second language. This was a replication of Lucy’s (1992) picture-matching task, in which participants were first shown a picture, and then had to choose a picture that is the most similar to the one previously shown out of several possibilities. The new pictures contained a change either in the number of animals, number of tools, or the amount of a substance presented in the scene. Intermediate English speakers, like Japanese monolinguals, were as likely to judge a picture as most similar when it included a change in amount of substance as in number of tools, whereas advanced English speakers gave more significance to a difference in amount of substance, matching the performance of English monolinguals.
The above experiment shows how learning a foreign language can move the boundaries of our conceptual categories, in this case by switching from a binary categorization system (animate vs. inanimate) to a ternary system (animate vs. inanimate-discrete vs. inanimate-continuous). An interesting extension of this line of research would be to study native English speakers learning Japanese and investigate whether their distinction between animate and inanimate objects would somehow become more delineated due to the selective use of grammatical number. We might find, for example, that English-Japanese bilinguals become quicker on a task in which they have to categorize an object as either animate or inanimate. Similarly, Japanese speakers have been found to primarily categorize objects based on their material, whereas English speakers tend to prefer categorization based on shape. This is also attributed to the lack of plural markings for inanimate objects in Japanese. Shape is usually the best indicator of object individuation (Athanasopoulos et al., 2008). Because of that, languages that encode individuation of inanimate objects through the use of plural markings (like English) draw the attention of their speakers to the shape of inanimate objects (Athanasopoulos et al., 2008). Athanasopoulos and colleagues tested this on Japanese-English bilinguals, using a task in which participants were presented with a novel shape of a specific colour. Then, they were asked to choose a shape that is “the same” as the one presented. The choice was between two stimuli: either a shape match, which had exactly the same shape but a different colour than the original, or a colour match, which had a different shape but the same colour. Even though all participants preferred a shape match over a colour match, this preference was stronger (88%) in advanced bilinguals than in intermediate bilinguals (69%) and approximated the scores of monolinguals speakers of English and Japanese (94% and 62%, respectively). These results suggest that learning a plural-marking language can induce object individuation, influencing the way we perceive the world and how we prefer to conceptualize it.
Grammatical aspect
Another grammatical feature that affects cognition is grammatical aspect. As discussed previously, speakers of aspect languages (like English) are more sensitive towards on-going actions and thus make use of an immediate reference frame when perceiving motion events (Bylund & Athanasopoulos, 2014; Bylund et al., 2013). This means that, for instance, when viewing a scene in which someone is cycling in a specific direction, getting closer to a house seen in the background, English speakers pay less attention to the end-point (i.e. the house) than speakers of a non-aspect language, and remember the event simply as
an on-going activity, i.e. someone is cycling. Bylund and Athanasopoulos (2014) investigated the effect of being a second-language speaker of English in cases where the native language is a non-aspect language. Because speakers of non-aspect languages pay more attention to the end-points when encoding motion events, the researchers were interested to see if the experience of learning a new, aspect language can influence their tendency to use the immediate reference frame. They tested native speakers of isiXhosa, a South African aspect-less language. All of the participants knew English and some spoke other aspect languages (Sesotho, Setswana, and SiSwati). However, the participants differed in the frequency of use of the aspect language(s), age of acquisition, and how much English schooling they had received in primary and secondary school. The task involved viewing three videos of motion events which either included reaching the end-point location (full goal orientation), did not include an obvious end-point (no goal orientation), or had an end-point somewhere in the background (intermediate goal orientation). The intermediate-goal video was the target and participants were asked to choose one from the other videos as being most similar to the target. Researchers found that the frequency of use of an aspect language as well as the amount of primary school English instruction inversely correlated with the preference for choosing the full goal orientation condition, meaning that the use of an aspect language had made the feature of ongoing-ness more salient for those speakers and had attenuated their native attentional preference towards goal orientation.
Motion events
Motion event encoding is also influenced by which feature of the act itself is more salient in the language being spoken. As I mentioned before, English puts more emphasis on the manner of motion than on its direction, expressing the manner by using an appropriate verb, such as jump or walk, and marking directionality by adding a preposition, such as in or out. Therefore, we would say that someone “jumped into the room” or “jumped out of the room”. Talmy (1991) has termed languages that primarily express manner in the main verb satellite-framed languages. By contrast, verb-framed languages (Spanish, Hebrew) typically express the direction of motion in the main verb. In Spanish, the above sentences would translate to “entró en la habitación saltando” and “salió de la habitación saltando” (literally “entered/exited the room jumping”). Hohenstein, Eisenber and Naigles (2006) devised an experiment in which they showed videos of motion events. Each event depicted a specific path (like entering or going uphill) and manner (like walking or running) of motion. The participants’ task was to watch the event and describe it either in Spanish or English, depending on the experimental condition. They then measured the frequency of usage of manner and path verbs when describing the event. The analysis showed, unsurprisingly, that participants used manner verbs more often when describing events in English than in Spanish. Interestingly, however, when the bilinguals were tested in Spanish, their pattern of responses did not resemble that of a monolingual Spanish population obtained in a previous experiment (Naigles, L. R., & Terrazas, 1998). The bilinguals used far more manner verbs than did the monolinguals, and did not use more path verbs than manner verbs. This result was especially visible for early bilinguals (age of English acquisition < 5 years old), who used more manner than path verbs even when describing the event in Spanish. These results suggest that a shift in event conceptualization has taken place in Spanish-English bilinguals, causing the manner of motion to become more salient when encoding motion events. The influence of the second language on the first was more pronounced than the influence of the first language on the second, meaning that the bilinguals differed more dramatically from the
Spanish-speaking monolingual population when tested in Spanish than they did from the English-Spanish-speaking monolingual population when tested in English (Hohenstein et al., 2006).
Grammatical tense / time perception
Grammatical aspect is one feature of verbal phrases through which languages exert influence on cognition. Another one is grammatical tense. In most Indo-European languages (like English, Spanish, or Russian), tense marking is an obligatory component of verbs. Therefore, a speaker of such languages must always express whether an action has taken place in the past, is happening in the present, or will happen in the future. However, some languages, like Mandarin or Indonesian, do not require marking verbs for tense, leaving the option to include temporal information at the discretion of their speakers or forcing the interlocutor to infer it from context (Boroditsky, Ham, & Ramscar, 2002). Presumably then—per the linguistic relativity hypothesis—speakers of tense-less languages will conceptualize temporal events differently than speakers of languages where tense is obligatory. Boroditsky and colleagues (2002) investigated this phenomenon on populations of monolingual speakers of English and Indonesian, and on Indonesian-English bilinguals. They collected pictures representing action events and arranged them in sets of three pictures each. Within one set, each of the pictures represented the same action performed by the same actor. The difference was that the action was depicted as happening either in the future, present, or the past—e.g. a man with his foot in the air about to kick a ball; a man with his foot touching a ball, as if kicking it; and a man with his foot sticking out and a ball in front of him, as if having just been kicked. In the first experiment, participants were shown two pictures which had different actors or different actions being performed. Their task was to rate the similarity of the two pictures. In the second, participants saw a picture representing one of the three tenses and then had to choose which one they had just seen out of the three options. The results showed that Indonesian speakers rated pictures depicting the same actor as more similar to each other than pictures depicting the same tense. English speakers’ results were exactly the opposite. English speakers were also better at remembering which picture they had seen than Indonesian speakers (41% vs. 31%, respectively). The Indonesian-English bilinguals’ results depended on the language they were tested in. Their performance was similar to English monolinguals when tested in English, but when tested in Indonesian, they judged the two conditions as equally similar, a result which was in between that of the two monolingual groups. For the second task, bilinguals reached 40% and 26% of correct responses when tested in English and Indonesian, respectively, matching the performance of monolingual speakers of the respective language.
These results show that lack of tense information in a language causes its speakers to pay less attention to that feature when encoding motion events. They also suggest that learning a language with obligatory tense markings can alter perception. In the case of the Indonesian-English bilinguals, tense became more salient, rendering the bilingual speakers more likely to encode it when perceiving a visual scene. This effect is not limited to second-language utterances, as the bilinguals’ performance was altered also when speaking Indonesian, showing that a permanent perceptual change had taken place, due to the influence of the second language.
How meaning affects perception in bilinguals
In the previous section, I have discussed how structural differences in the grammar of different languages give rise to different ways of perceiving and conceptualizing the world in speakers of multiple languages.
Next, I will look at semantic differences and how they influence where we place borders between conceptual categories.
Conceptual categories
Translating between languages is a difficult task partly because a concept expressed in language A often does not have a direct equivalent in language B. This is true even for languages very closely related to each other: Ukrainian has one word стіна́ (stina) referring to both an interior and an exterior wall, while neighbouring Polish has two separate words for them (ściana and mur, respectively). A related Ukrainian word перегоро́дка (peregorodka) refers to an interior partition wall that is not load-bearing and стіна́ is used for one that is load-bearing, whereas Polish uses ściana for both. This example shows that even the smallest, seemingly insignificant details in the external world can be a basis for differentiation between two concepts. If the linguistic relativity hypothesis is true, it follows that speakers of different languages should look at the same scene in a different way, placing the same objects in different conceptual categories.
What happens, then, when a bilingual speaker enters into a room and creates a mental representation of the objects contained therein? Malt and Sloman (2003) created an experiment in which they used sets of objects which could be described using several labels, obtained originally from a native English-speaking population. For example, one set of stimuli contained pictures of jars, containers, and bottles. They then presented those pictures to second-language English speakers of varying degree of fluency and immersion in an English-speaking environment. The task was to name each of the pictures using own words. The experimenters then compared how much overlap occurred between the categories used by the native population and the experimental groups. The results showed that even the most advanced second-language English speakers differed significantly from the monolinguals. However, fluency in English correlated with overlap in naming patterns. The best predictor for performance was the time spent immersed in an English-speaking environment. This suggests that the native language exerts a strong influence on conceptual categories, even after several years of formal instruction and immersion in a second-language environment. Bilingual speakers, once they become highly fluent, exhibit performance that differs both from beginners and from monolinguals. Unfortunately, native English speakers mastering a second language were not included in the study, therefore it is difficult to assess the degree of second-language influence on category naming.
Spatial metaphors of time
Language can also affect cognition through less obvious means, such as metaphors contained within it. One such example is metaphors that people use to talk about time, specifically how they represent time in space. Representing time as space is common across different cultures and languages but different languages employ different metaphors in doing so (Boroditsky, Fuhrman, & McCormick, 2011). In English, as in many other languages, it is common to represent time horizontally, placing earlier events in front and later events in the back. We might say, for example, that we are “looking forward to an upcoming party” or “thinking back to last weekend”. Although other ways of spatial positioning in time metaphors exist in English, the vast majority of them relate to the notion of horizontal time. In Mandarin, even though horizontal time metaphors are dominant as well, it is also common to talk about time in a vertical way (Boroditsky, 2001). Mandarin speakers say that preceding events are up, whereas upcoming events are
down, and they use these metaphors when talking about the previous and next months. Up to a third of all spatial metaphors of time investigated by Chen (2007) involved talking about time vertically, showing that this usage occurs frequently in Mandarin.
Boroditsky and colleagues (2011) have reviewed evidence showing that this difference in speaking about time results in different ways of thinking about time and, thus, in different patterns of behaviour in cognitive tasks among speakers of English and Mandarin. Furthermore, they devised a task in which they showed two pictures with a clear relation of time between them (e.g. a picture of a young boy and one of an old man). Participants had to indicate, by pressing an appropriate button, if the first picture was taken earlier or later than the second one. The experimenters balanced the position of the buttons, so that for some participants they were arranged vertically and for others, horizontally. They found that English participants responded faster when the keys were arranged horizontally than when they were arranged vertically. Furthermore, they were faster if the left key corresponded to an answer “earlier” and the right to “later”, than when the keys were reversed. No such effect was found for the vertical arrangement. On the other hand, Mandarin-English bilinguals (for whom Mandarin was the first language) responded as quickly when the keys were arranged vertically as when they were arranged horizontally. They benefited both from arranging the keys congruently with the time metaphors on the horizontal axis (i.e. left is earlier) and on the vertical axis (i.e. up is earlier).
This experiment is somewhat problematic since no monolingual Mandarin-speaking population was included and, as I have discussed before, the second language of the bilingual speakers will also have affected their performance. However, what remains clear is that Mandarin speakers do perceive time in a vertical way, whereas English speakers do not. This result shows that concepts deeply rooted in the semantics of a language through metaphors can influence the way that speakers conceptualize the world.
Mental concepts in bilinguals
In the previous sections, I have demonstrated and discussed evidence for how speaking two languages affects the way we perceive and categorize the world. All the discussed examples were concerned with micro-effects of bilingualism, that is, alterations in cognitive or perceptual processes arising due to the grammatical or conceptual structure of a particular language, or due to the difference in those structures found between a pair of languages. In some cases, bilinguals’ performance was matching that of monolingual speakers of one of the languages and in others, they situated themselves in between the two groups. These discrepancies have led different researchers to reach different conclusions about how mental concepts are organised in the bilingual mind. In this section I will present a theoretical framework put forward by Bassetti and Cook (2011), that aims to abstract from the empirical data and form a more general theory of how mental concepts function in bilinguals in different situations.
Bassetti and Cook (2011) describe four possibilities for the relationship between language (specifically, language’s vocabulary) and cognition. In the one-concept scenario, a single mental concept is attached to two labels, each corresponding to a single language. The concept comes from one of the languages and is directly accessed when speaking the other one, through the attached label. For example, an English-Italian speaker might have one concept of “blue”, coming from English, which would be invoked when speaking Italian both by talking about blu (Italian for “dark blue”) and azzurro (Italian for “light blue”). This way of
structuring mental concepts is common for beginner second language learners who still rely on their first-language concepts and directly map newly learnt words to them (Bassetti & Cook, 2011). However, this scenario is not exclusively observed in beginners. As we have seen in the previous sections, intermediate foreign language speakers perform similarly to monolinguals of their first language when it comes to number perception (Athanasopoulos, 2006), suggesting that they might rely on mental concepts acquired through their first language. Advanced speakers in the same study, on the other hand, performed like second-language monolinguals. Possibly, a restructuring of the mental repertoire takes place after acquiring fluency in the second language, and consequently the mental concept shifts to the corresponding second-language construct. In other domains it might be possible to retain the original construct even after acquiring a high level of fluency, as bilinguals have been shown to still make use of their first language concepts when speaking in the second language and to label common objects differently than monolinguals, even after many years spent abroad (Malt & Sloman, 2003).
A second possibility, the double-concept scenario, describes a situation in which a bilingual speaker has two separate concepts, each corresponding to one language. Depending on which language is being spoken, the corresponding concept is activated. In the case of colours, this implies that an Italian-English bilingual would activate a concept of “dark blue” when talking about blu in Italian, but when talking about blue in English, a lighter shade of blue would be conceptualised. As speakers become more proficient in a second language, differences between the two labels might become more salient, leading the underlying representation to split into two interconnected, single-language mental concepts (Bassetti & Cook, 2011). In this scenario one would expect the language in which participants are being tested to affect the obtained results. This is indeed the case at least in several domains. When processing action events, as I have discussed before, Indonesian-English bilinguals are poor at remembering if the action they had just seen happened in the past, present, or future (Boroditsky et al., 2002). This effect is only present when tested in Indonesian—a language which does not require specifying tense. When tested in English—a language in which tense marking is obligatory—their performance is significantly higher. Similar results have also been obtained for processing the emotional content of words (Sutton, Altarriba, Gianico, & Basnight-Brown, 2007). Furthermore, one could argue that in the experiment on number perception mentioned above (Athanasopoulos, 2006), advanced bilinguals have shifted from a one-concept representation, derived from the first language, to a double-concept representation. Therefore, when tested in their second language (as was the case in this experiment), their performance approached that of monolingual speakers of that language. However, more research involving testing in both languages is needed in this domain. Nevertheless, these examples suggest that, at least in some cases, bilinguals think differently depending on which linguistic environment is the current context.
A third possibility, the one-integrated-concept scenario describes a situation in which the two concepts have merged into a single one which integrates features of both. In the case of colour perception, a bilingual speaker will think of a single notion of “blue” which will be darker than the English blue but lighter than the Italian blu. In this case, bilinguals’ conceptual space would differ from that of monolingual speakers of either language, placing them somewhere in between the two languages. Much empirical data supports this notion. In the domain of gender, as we have seen before, second-language learners seem to increasingly shift their mental attribution of gender in their first language, resembling native speakers of their second language more and more as they become more fluent (Kurinski & Sera, 2011).
When encoding motion events, bilingual speakers of English and Spanish seem to conceptualize them not primarily in terms of manner—as English speakers do—nor dominantly in terms of path—as Spanish speakers do. Instead, they are more likely to primarily encode the path of the event than English speakers, and more likely to primarily encode the manner than Spanish speakers. This becomes especially evident when tested in Spanish (Hohenstein et al., 2006). In the domain of time perception, the aforementioned Indonesian-English bilinguals did exhibit performance laying in-between both monolingual groups when tested in Indonesian (Boroditsky et al., 2002). The results could therefore also be interpreted as evidence for the one-integrated-concept scenario, as they suggest that at least the concept associated with the Indonesian label has become integrated with that connected to English. Similar results for integrated concepts in bilingual speakers have been obtained for motion conceptualization, temporal representations, and concepts of animals (Bassetti & Cook, 2011).
A final, fourth possibility described by Bassetti and Cook (2011), the original-concept scenario, refers to a situation in which the bilingual speaker represents a concept in a new, different way from that of either of his or her languages. This concept is, furthermore, not merely a merged or in-between version of the two concepts, as in the scenario described above, but rather, it is an entirely new, original category. In the case of colour perception, the bilingual speakers could represent “blue” as a different shade of blue than either English blue or Italian blu. Some supporting evidence comes from the research of Athanasopoulos (2009), who showed that Greek-English bilinguals develop a concept of “light blue” that lies not between the lighter Greek ghalazio and the darker English blue (as would be expected from the integrated-concept scenario) but is in fact an even lighter shade of blue.
When thinking about this framework, it is important to note that different scenarios can occur simultaneously within the conceptual repertoire of an individual speaker and every single concept could be represented in a specific way. It is entirely possible that distinct types of concepts are encoded in distinct ways, which would explain why different researchers, having examined different aspects of cognition, have provided support for different scenarios of conceptual representation in bilinguals. In fact, there exists some empirical evidence showing that the same people have several varieties of concepts even within the same domain, such as developing a new concept for “orange” but preserving the concepts of “blue” and “green” from one of the languages (Jameson & Alvarado, 2003).
All in all, the exact conceptual representation that a bilingual speaker will develop seems to depend on several factors. The first is the level of experience with the second language. As speakers become more fluent, they tend to shift from purely one-concept representations, based on the first language, to representations involving double or integrated concepts. Pure bilinguals, fluent in both of their languages, might be more likely to develop a single integrated concept based upon their constant experience with labels in both languages.
The second factor is the linguistic domain. In certain areas, such as categorizing common objects and attributing gender to objects, the first language seems to exert more influence on a bilingual’s conceptual space than the second language, leading to representations that align more with the one-concept scenario. In other areas, like time perception and motion events, it seems to be easier to decouple the two languages, leading to representations resembling the double-concept and one-integrated-concept scenarios. It is worth noting the possibility that these distinctions arise merely as a result of specific
experimental strategies used by researchers investigating these domains. More research, specifically aimed at distinguishing between the different possibilities of mental concept arrangement, will shed more light on this phenomenon, allowing to form an explanation regarding why specific domains allow for specific types of representations.
Another potentially relevant factor is the linguistic history of particular speakers. Simultaneous bilinguals, raised with both languages from birth, will be less likely to have a dominant language upon which the mental concepts could be based and will thus have more integrated and decoupled concepts than sequential bilinguals, for whom the first language has been dominant for at least the first years of their lives.
Finally, a fourth plausible factor is the conceptual overlap between two languages. It could be that only specific linguistic arrangements warrant forming an original concept, as is the case with the words for “blue” in the Greek-English language pair. The larger the overlap, I predict, the more likely a speaker will be to develop a single integrated concept or to keep relying on a single concept from one of the languages. In situations of less overlap between the two concepts, on the other hand, I predict a higher prevalence of double concepts and original concepts. However, research specifically designed to test these predictions is lacking at the moment.
The factors I have outlined correspond roughly to the taxonomy presented by Pavlenko (2005). She describes how conceptual changes are influenced by language, and distinguish between individual, interactional, and psycholinguistic factors. Individual factors include the language learning history, language proficiency, the degree of acculturation, and domain expertise of the speaker. Interactional factors include the context, and linguistic proficiency of the interlocutor. Psycholinguistic factors relate to the specific languages spoken by the bilingual speakers, and include the comparability of concepts in the two languages in question, as well as how the concept is encoded in each language (Pavlenko, 2005). Quite possibly, there exist other factors determining the mental concept scenario applied for a particular conceptual domain by a particular speaker. Future research should keep this framework in mind when designing experimental paradigms, in order to make it possible to distinguish between different scenarios.
Discussion
I have reviewed some of the evidence relating the linguistic relativity hypothesis and bilingualism. The studies I discussed show that linguistic relativity can be observed in bilingual speakers and studies should focus on these populations to investigate how language affects cognition. I will now briefly present some concerns that can be raised regarding the linguistic relativity hypothesis and discuss whether they threaten the issue at hand. Then I will summarize the findings presented in this review and sketch the emerging picture of the bilingual mind.
Alternative explanations
One objection that can be raised about the observed effects, is that they might not reflect how the mind works in general but merely showcase cognitive habits that people acquire due to the influence of the language they speak. In the domain of time perception, as I have discussed before, Mandarin speakers are more likely to conceptualize time in a vertical way, placing past events at the top and future events at
the bottom (Boroditsky et al., 2011). However, in one study (Boroditsky, 2001), the researchers first trained native English speakers to use the vertical time metaphor. They then found that their results after training looked more like those of native Mandarin speakers. It seems as if the short training was enough to cause the participants to make use of this new way of representing time. Given this result, one might argue that the difference between English and Mandarin speakers, with regard to representing time, does not have a perceptual basis. Rather, Mandarin speakers habitually tend to represent time vertically on some occasions and English speakers can do the same should they see it fit or beneficial in a particular situation, as was the case after having been trained to do so.
The interpretation that, at least in some cases, effects of language are temporary, is in line with the thinking-for-speaking hypothesis put forward by Slobin (1987). He proposed a weaker version of linguistic relativity, claiming that language affects cognition during the process of language use. In this view, the thought process is language-specific and related to the specific content that is being produced (de Groot, 2013). Thinking-for-speaking would be a distinct mode of cognition in which “one unconsciously focuses on those aspects of objects and events that are most readily encodable in one’s particular language” (Slobin, 2003, p. 157). According to this approach, constantly speaking a certain language creates a habit of attending to particular details of the environment or of interpreting it in a particular way (Tajima & Duffield, 2012). However, thinking-for-speaking is in fact very much in accordance with the framework presented here, as Slobin himself argued that the effects of thinking-for-speaking can lead to the development of specific attentional patterns which, in turn, require the speaker to form a mental representation available to be used even in the absence of language (Slobin, 2003). Similarly, with regard to the example above, the fact that it is possible for English speakers to think about time vertically does not mean that the metaphors contained in the English language do not shape the way that English speakers conceptualize temporal events, whether by habit or not.
Another concern is the possibility that, in the studies of cognitive effects of language, the obtained results are not due to language learning but, instead, come about because of verbalization during task performance. This problem is difficult to address empirically. Even though researchers use several methods to prevent verbalization, such as explicit instructions or filler tasks, it cannot be ruled out that participants are using an internal monologue to help themselves in performing the task. In the domain of grammatical gender we have seen that speakers of non-gender languages begin to nevertheless assign gender to objects as they become more fluent in the second language which contains grammatical gender. These assignments, as I have discussed previously, align with the grammatical gender of that object in the second language, even when participants are tested in their first language. The concern here is that this effect may not reflect a change in the mental concept of the associated object but rather arise as a result of internally vocalizing the name of the object in the second language. This could also explain why, in certain experiments, bilingual speakers were performing differently depending on the language probed. The linguistic relativity hypothesis has also been challenged by researchers claiming that the described effects are the result of culture, rather than language (see Masuda & Nisbett, 2001, for an example of such an explanation; and Tajima & Duffield, 2012, for a rebutal from a linguistic relativity perspective). This claim, however, can be addressed with carefully designed studies including bilingual populations. Bassetti (2007) investigated gender perception in Italian-German bilinguals and Italian monolinguals.
Interestingly, both groups were native speakers of Italian and were living in the same town in Italy, differing only by the school they attended (one of them had German as the language of instruction). This design allowed the researcher to conclude that the differences she had found were solely the result of language affecting cognition.
All in all, the amount of evidence for linguistic relativity outweighs the few minor concerns that can be raised against it. However, there are several challenges that face the field at the moment. Many studies I discussed here involved subjective measures, such as similarity judgments. While some conclusions can certainly be derived from such “soft” measurements, researchers have to be more careful with their interpretations when using those subjective scales and supplement their findings with results from other experiments that measure the phenomenon more objectively, such as using reaction time. Furthermore, all of the studies I discussed here are purely behavioural. It is only in recent years that researchers have begun to approach linguistic relativity from a cognitive neuroscience perspective, looking for brain activity related to cognitive changes resulting from language (see e.g. Boutonnet, Dering, Viñas-Guasch, & Thierry, 2013; Flecken, Athanasopoulos, Kuipers, & Thierry, 2015). More research in this vein is presently needed in order to further advance our understanding of linguistic relativity.
The bilingual mind
The empirical research and the theoretical frameworks I have outlined here present a picture of the bilingual mind that is simultaneously complex and inviting for future research. Bilinguals of the world— which constitute the majority of language speakers (Bassetti, 2007)—have some common traits, such as being good at inhibiting conflicting information or having a larger conceptual repertoire, which enables them to be more creative and better at task switching. However, once specific language pairs and their effects are taken into account, a vast range of variability is revealed. Mental concepts can be arranged in several different ways, depending presumably on factors such as which grammatical features are involved or the level of experience with each language spoken, and presently more research is needed to understand the peculiarities of the conceptual repertoire of a bilingual speaker. Effects of language on various aspects of cognition are not clear-cut either—depending on the linguistic domain studied and the empirical method used, studies have either shown bilinguals to approximate the results of monolingual speakers, situated them in-between the two populations, or shown them to think or perceive in a completely different way than monolingual speakers do.
With this paper I did not intend to resolve these discrepancies but rather summarize the current findings and direct attention towards the need for new experimental designs, more theoretically-informed and targeted at untangling the various possibilities of mental space arrangement in bilingual speakers. I have presented several scenarios of how mental concepts could be represented in a bilingual mind, put forward by Bassetti and Cook (2011). Future research should focus on arbitrating between the different scenarios by incorporating this theoretical framework in designing the experiments.
This paper also attempted to show the importance of bilingualism in the further investigation of linguistic relativity and its implications. Bilingual speakers constitute a major part of the world population, yet for a long time, research on linguistic relativity (and in cognitive linguistics in general) has been ignoring that fact and focusing on monolingual speakers, considering them a norm to which bilinguals’ performance
was being compared (de Groot, 2011). The tendencies are now shifting, and in the past decade researchers have begun to realize the value and importance of including bilinguals into their research designs (Athanasopoulos & Aveledo, 2013). When it comes to linguistic relativity, one cannot think of a better population to study this hypothesis on. Researchers can minimize the effect of sociocultural or geopolitical differences by studying bilingual and monolingual populations living in the same country or city, brought up and abiding by the same system of values, religion, and so on. Researchers can also treat the language in which the experiment is conducted as the independent variable and give the same task to the same person twice, thus minimizing the effect of individual variance on their results.
Furthermore, I organized the findings discussed in this paper according to the framework put forward by Gentner and Goldin-Meadow (2003), who describe effects of language on cognition as targeting the perception of the external world (“lens”), the categorization of objects into mental concepts (“category-maker”), and the way of thinking and solving problems (“toolkit”). The benefit of this approach is in segregating the findings and evaluating the conclusions made by researchers based on which aspect of cognition is affected by the language studied. I have shown that grammatical differences primarily affect the lens of language, while semantic difference influence the category-maker.
I discussed the effects of language on cognition in several linguistic domains, showing that these effects can arise from differences in grammatical structures, as well as from different patterns of categorization and assigning meaning in various languages. Various aspects of grammar affect perception in distinct ways, in some cases leading to different results. I have shown how the presence of grammatical gender in the one language affects perception of objects in the other, causing the bilingual speakers to perceive objects as more masculine or feminine. In the field of grammatical number, I have shown how learning a second language can change the way speakers perceive inanimate objects and causes them to develop a new categorical distinction between inanimate-discrete and inanimate-continuous objects. Furthermore, the use of grammatical number in a language causes their speakers to selectively pay more attention to the shape or the material of the object at hand. I also discussed grammatical aspect, and how learning a language with grammatical aspect causes bilinguals to encode the end points of motion events less preferentially, leading, in turn, to increased salience of the ongoing-ness aspect of an action. I have also shown that the grammatical construction of a verbal phrase describing motion events influences whether the speaker preferentially encodes the direction or the manner of movement, and that bilinguals’ responses in their first language were influenced by their second language in this domain. Finally, I discussed evidence showing that, for speakers whose first language does not require grammatical tense, learning a second language (one which does require it) leads to an increase in salience of temporal information when encoding a scene.
Aside from the grammatical aspect of language, I have looked at how categorization is affected by semantic differences in attaching meaning to various objects between languages. I have shown that the first language exerts a strong influence on this aspect of cognition, showing evidence that people still make use of their first-language categories for common-use objects even after many years spent living abroad. Furthermore, I discussed how the use of linguistic tools such as metaphors in the second language can change the way bilinguals conceptualize time.
The picture that emerges from the studies I reviewed is one that cannot be summarized by a “one size fits all” theory of the mental representation of the external world in the bilingual mind. Rather, several possibilities of structuring mental concepts have to be accepted to take into account the experimental data. Using the Bassetti and Cook’s (2011) framework, I made an attempt at categorizing empirical results as providing support for one of the scenarios according to which mental concepts can be represented. Evidence from the domain of grammatical gender and common-use object categorization seems to point to the one-concept scenario, in which both linguistic labels in the two languages are connected to the same mental concept, coming from the first language. Results of intermediate second language speakers in the domain of grammatical number also provide evidence for this scenario. On the other hand, the results of advanced speakers in this domain, like the results from the field of motion perception, and motion events, seem to support the double-concept scenario, in which a bilingual speaker has two mental concepts, each connected to a linguistic label in one of their languages. The one-integrated-concept scenario describes a merged concept with elements from both languages. Some results that I have discussed agree with this scenario, such as the increasing likelihood of ascribing gender to objects in correlation with the length of study of a gender language. The evidence from the domain of time perception could also be interpreted as providing support for this scenario. In discussing the Bassetti and Cook (2011) framework of mental concepts, I have further attempted to isolate factors that might be responsible for the emergence of a specific type of scenario, such as fluency and language acquisition history, linguistic domain, and the conceptual overlap between the two languages. Currently, more research is needed to ascertain which of these factors play the most important role in mental concept development and to distinguish other causes of developing a specific scenario.
Conclusion
Linguistic relativity, even though it had been disregarded historically, has taken its firm place in language and cognition research and is generally embraced within cognitive sciences. With the work I presented here, I attempted to illustrate the importance and usefulness of this notion in understanding the human mind. Spoken or written language is a primary means of communication between humans, serving to introduce ideas, solve problems, and engage in discussions. Unsurprisingly then, it has been extensively researched in various fields, such as linguistic, artificial intelligence, and cultural studies. Cognitive science attempts to explain how our minds deal with language and provide a neural basis for linguistic activities. Even though some aspects of how we acquire, process, understand, and produce language are currently still not well understood, current psychological approaches and neuroscientific methods are moving us closer towards a complete theory of a linguistic mind. By focusing on bilingual populations when approaching this problem, researchers will benefit greatly due to limiting the influence of confounding group factors such as culture, geography, or history. Individual intra-speaker variability also plays a less important role when the same individual can serve as subject in both of the studied languages. Finally, studying bilinguals to answer questions regarding linguistic relativity will yield results of higher relevance and applicability in a world where bilingualism is the norm.
The bilingual mind, as discussed in this paper, is proving to have a very complex conceptual organization, showing behaviour vastly different than that of monolingual speakers. Very likely, bilingual speakers are capable of creating new concepts by integrating those found in other languages or by extending the
concepts from one language to the other. Clearly, a bilingual speaker is not just someone mapping newly heard linguistic labels onto existing ideas. Instead, the bilingual mind should be seen as a dynamic system of interconnected representations and multiple ways of conceptual arrangement.
References
Athanasopoulos, P. (2006). Effects of the grammatical representation of number on cognition in bilinguals. Bilingualism: Language and Cognition.
Athanasopoulos, P. (2009). Cognitive representation of colour in bilinguals: The case of Greek blues *. Bilingualism: Language and Cognition, 12(1), 83–95. https://doi.org/10.1017/S136672890800388X Athanasopoulos, P., & Aveledo, F. (2013). Linguistic relativity and bilingualism. Memory, Language, and
Bilingualism: Theoretical and Applied Approaches., 236–255.
Athanasopoulos, P., Kasai, C., Ameel, E., Storms, G., Malt, B. C., Sloman, S. A., … Gelman, R. (2008). Language and thought in bilinguals: The case of grammatical number and nonverbal classification preferences. Applied Psycholinguistics, 29(1), 105–123. https://doi.org/10.1017/S0142716408080053
Bassetti, B. (2007). Bilingualism and thought: Grammatical gender and concepts of objects in Italian-German bilingual children. International Journal of Bilingualism, 11(3), 251–273. https://doi.org/10.1177/13670069070110030101
Bassetti, B., & Cook, V. (2011). Relating language and cognition: The second language user. In Language and bilingual cognition.
Bialystok, E. (2005). Consequences of bilingualism for cognitive development. Handbook of Bilingualism. Boroditsky, L. (2001). Does Language Shape Thought?: Mandarin and English Speakers’ Conceptions of
Time. Cognitive Psychology, 43, 1–22. https://doi.org/10.1006
Boroditsky, L., Fuhrman, O., & McCormick, K. (2011). Do English and Mandarin speakers think about time differently? Cognition, 118(1), 123–129. https://doi.org/10.1016/j.cognition.2010.09.010
Boroditsky, L., Ham, W., & Ramscar, M. (2002). What is universal in event perception? Comparing English and Indonesian speakers. Proceedings of the 24th …, 6.
Boutonnet, B., Dering, B., Viñas-Guasch, N., & Thierry, G. (2013). Seeing Objects through the Language Glass. Journal of Cognitive Neuroscience, 25(10), 1702–1710. https://doi.org/10.1162/jocn_a_00415 Bylund, E., & Athanasopoulos, P. (2014). Language and thought in a multilingual context: The case of isiXhosa. Bilingualism: Language and Cognition, 17(2), 431–441. https://doi.org/10.1017/S1366728913000503
Bylund, E., Athanasopoulos, P., & Oostendorp, M. (2013). Motion event cognition and grammatical aspect: Evidence from Afrikaans. Linguistics, 51(5), 929–955. https://doi.org/10.1515/ling-2013-0033 Chen, J.-Y. (2007). Do Chinese and English speakers think about time differently? Failure of replicating
Boroditsky (2001). Cognition, 104(2), 427–436. https://doi.org/10.1016/j.cognition.2006.09.012 Cook, V., & Bassetti, B. (2011). Language and bilingual cognition.
de Groot, A. M. B. (2011). Cognitive consequences of bilingualism and multilingualism. In Language and cognition in bilinguals and multilinguals: An introduction (pp. 339–404). NY: Psychology Press. de Groot, A. M. B. (2013). About doublespeak and doublethink. The American Journal of Psychology,
Devitt, M., & Sterelny, K. (1999). Language and Reality: an Introduction to the Philosophy of Language. Proceedings and Addresses of The American. Oxford: Blackwell. https://doi.org/10.1353/lan.2001.0002
Flecken, M., Athanasopoulos, P., Kuipers, J. R., & Thierry, G. (2015). On the road to somewhere: Brain potentials reflect language effects on motion event perception. Cognition, 141, 41–51. https://doi.org/10.1016/j.cognition.2015.04.006
Foundalis, H. (2002). Evolution of gender in Indo-European languages. Proceedings of the 24th Annual Conference.
Gentner, D., & Goldin-Meadow, S. (2003). Language in mind: Advances in the study of language and thought. Massachusetts Institute of Technology.
Hohenstein, J., Eisenberg, A., & Naigles, L. (2006). Is he floating across or crossing afloat? Cross-influence of L1 and L2 in Spanish–English bilingual adults. Bilingualism, 9(3), 249. https://doi.org/10.1017/S1366728906002616
Jameson, K., & Alvarado, N. (2003). Differences in color naming and color salience in Vietnamese and English. Color Research & Application.
Kousta, S., Vinson, D., & Vigliocco, G. (2008). Investigating linguistic relativity through bilingualism: The case of grammatical gender. Journal of Experimental.
Kurinski, E., & Sera, M. D. (2011). Does learning Spanish grammatical gender change English-speaking adults’ categorization of inanimate objects? Bilingualism: Language and Cognition, 14(2), 203–220. https://doi.org/10.1017/S1366728910000179
Lewis, M., Simons, G., & Fennig, C. (2009). Ethnologue: Languages of the world.
Lucy, J. (1992). Language diversity and thought: A reformulation of the linguistic relativity hypothesis. Malt, B. C., & Sloman, S. a. (2003). Linguistic diversity and object naming by non-native speakers of English.
Bilingualism: Language and Cognition, 6(1), 47–67. https://doi.org/10.1017/S1366728903001020 Marlowe, F. (2010). The Hadza: hunter-gatherers of Tanzania.
Masuda, T., & Nisbett, R. (2001). Attending holistically versus analytically: Comparing the context sensitivity of Japanese and Americans. Journal of Personality and Social.
Naigles, L. R., & Terrazas, P. (1998). Motion-verb generalizations in English and Spanish: Influences of language and syntax. Psychological Science, 9(5), 363–369.
Pavlenko, A. (2005). Bilingualism and Thought. In J. F. Kroll & A. M. B. De Groot (Eds.), Handbook of bilingualism Psycholinguistic approaches (Vol. 29, pp. 9–29). https://doi.org/10.1017/S0272263107210071
Phillips, W., & Boroditsky, L. (2003). Can quirks of grammar affect the way you think? Grammatical gender and object concepts. Proceedings of the 25th Annual Meeting of the.
Pinker, S. (1994). The Language Instinct: The New Science of Language and Mind. Penguin Books. https://doi.org/10.1002/cne.22311
Scheutz, M., & Eberhard, K. (2004). Effects of morphosyntactic gender features in bilingual language processing. Cognitive Science.
Sera, M., Berge, C., & Pintado, J. del C. (1994). Grammatical and conceptual forces in the attribution of gender by English and Spanish speakers. Cognitive Development.
Sera, M., Elieff, C., Forbes, J., & Burch, M. (2002). When language affects cognition and when it does not: An analysis of grammatical gender and classification. Journal of.
Slobin, D. (1987). Thinking for speaking.
Slobin, D. (1996). Two ways to travel: Verbs of motion in English and Spanish. Grammatical Constructions: Their Form and Meaning.
Slobin, D. (2003). Language and thought online: Cognitive consequences of linguistic relativity. Language in Mind: Advances in the Study of Language.
Sutton, T. M., Altarriba, J., Gianico, J. L., & Basnight-Brown, D. M. (2007). The automatic access of emotion: Emotional Stroop effects in Spanish-English bilingual speakers. Cognition Emotion, 21(5), 1077–1090. https://doi.org/10.1080/02699930601054133
Tajima, Y., & Duffield, N. (2012). Linguistic versus cultural relativity: On Japanese-Chinese differences in picture description and recall, 23(4), 675–709. https://doi.org/10.1515/cog-2012-0021
Talmy, L. (1991). Path to realization: A typology of event conflation. Annual Meeting of the Berkeley Linguistics.
