Notes from Aneta Pavlenko’s The Bilingual Mind

The book surveys what we know about the relationship between language and thought. In addition to outlining and critiquing the historical development of the field of study, Pavlenko rigorously reviews and summarizes the current literature, focusing on experimental studies, particularly those done with bilingual participants.

Table of Contents §

1. The Sapir-Whorf Hypothesis: A Story of Manufactured Consent
   1. Wilhelm von Humboldt (1767 – 1835)
   2. Franz Boas (1858 – 1942)
   3. Edward Sapir (1884 – 1939)
   4. Benjamin Lee Whorf (1897 – 1941)
   5. The Real Authors of the Sapir-Whorf Hypothesis
2. Bilingualism: Key Terms and Constructs
3. The Evolution of Language
   1. The Evolution of Symbolic Cognition
   2. The Development of the Prehistoric Mind
   3. Language, Cognition, and Linguistic Thought
4. Linguistic Diversity and Language Change
5. Categorical Perception
   1. Referential Indeterminacy
6. Color
   1. “Basic Color Terms” Theory
   2. (Lack of) Color Terms in Homeric Greek
   3. Language Effects on Perception of Color
   4. Bilinguals’ Color Categories
7. Literature

The Sapir-Whorf Hypothesis: A Story of Manufactured Consent §

The so-called “Sapir-Whorf hypothesis” was never postulated by either of its two namesakes — it was a radical restructuring of their ideas by experimental psychologists. This resulted in a loss of nuance and complexity and a shift from the idea of “habitual thought”.

Wilhelm von Humboldt (1767 – 1835) §

• “Weltansicht” — the fundamental capacity of the mind to process the world through language and to organize it into concepts
• “Weltanschauung” — an interpretive system or ideology that is subjective and not language-bound
• “Underhill (2009) emphasizes that the cornerstone of his linguistic philosophy was Weltansicht, the largely unconscious way in which we follow the patterns of our language in negotiating daily reality. For Humboldt, these patterns were neither predetermined nor arbitrary, rather they reflected the ongoing mental activity, in which language, an interactive tool of human cognition, accommodated our evolving needs, and the relationship between the mind, the language, and the world was dynamic and mutually constitutive.” ​(p. 3)

Franz Boas (1858 – 1942) §

• “[Boas] chief interest was in cross-linguistic variation in obligatory categories — that is, in what the different languages require you to encode — and its implications for mental activities.” ​(p. 8)
• “he was intrigued by the ways in which the automatic nature of language use placed these differences below the threshold of awareness, making the categories of one’s native language appear ‘objective’.” ​(p. 8)

Edward Sapir (1884 – 1939) §

Such categories as number, gender, case, tense, mode, voice, ‘aspect’ … are not so much discovered in experience as imposed upon it because of the tyrannical hold that linguistic form has upon our orientation in the world.

Sapir, [1931] 1964: 128, via Pavlenko, 2014: 8

Language is a guide to ‘social reality’. Though language is not ordinarily thought of as of essential interest to the students of social science, it powerfully conditions all our thinking about social problems and processes. Human beings do not live in the objective world alone, nor alone in the world of social activity as ordinarily understood, but are very much at the mercy of the particular language which has become the medium of expression for their society. It is quite an illusion to imagine that one adjusts to reality essentially without the use of language and that language is merely an incidental means of solving specific problems of communication or reflection. The fact of the matter is that the ‘real world’ is to a large extent unconsciously built up on the language habits of the group. No two languages are ever sufficiently similar to be considered as representing the same social reality. The worlds in which different societies live are distinct worlds, not merely the same world with different labels attached … We see and hear and otherwise experience very largely as we do because the language habits of our community predispose certain choices of interpretation.

Sapir, [1929] 1949: 162, via Pavlenko, 2014: 9

• “‘the tyrannical hold’, does not imply an actual constraint on thought — language patterns, for Sapir, are akin to grooves which may be easier to follow automatically yet may be overcome, through poetic expression, linguistic study, or the process of learning a foreign language.” ​(p. 9)

To pass from one language to another is psychologically parallel to passing from one geometrical system of reference to another. The environing world which is referred to is the same for either language; the world of points is the same in either frame of reference. But the formal method of approach to the expressed item of experience, as to the given point of space, is so different that the resulting feeling of orientation can be the same neither in the two languages nor in the two frames of reference. Entirely distinct, or at least measurably distinct, formal adjustments have to be made and these differences have their psychological correlates.

Sapir, [1924] 1949: 153, via Pavlenko, 2014: 9

Benjamin Lee Whorf (1897 – 1941) §

Users of markedly different grammars are pointed by their grammars towards different types of observations and different evaluations of externally similar acts of observation, and hence are not equivalent as observers but must arrive at somewhat different views of the world.

Whorf, [1940] 2012: 282–283, via Pavlenko, 2014: 10

• “Like Humboldt and Sapir before him, Whorf, too, believed in the plasticity of the human mind and its ability to go beyond the categories of the mother tongue.” ​(p. 10)

The Real Authors of the Sapir-Whorf Hypothesis §

• “it had never occurred to Sapir and Whorf to put forth testable hypotheses. […] neither [Whorf] nor Sapir ever elaborated the meaning of ‘different observations’ or ‘psychological correlates’” ​(p. 12)
• “The term ‘the Sapir-Whorf hypothesis’ was first used by linguistic anthropologist Harry Hoijer (1954b) to refer to the idea ‘that language functions, not simply as a device for reporting experience, but also, and more significantly, as a way of defining experience for its speakers’ (p. 93)” ​(p. 13)
• “This version […] was soon replaced by an alternative, developed by psychologists Roger Brown and Eric Lenneberg, who translated Sapir’s and Whorf’s ideas into two ‘testable’ hypotheses.” ​(p. 13)

Linguistic relativity holds that where there are differences of language there will also be differences of thought, that language and thought covary. Determinism goes beyond this to require that the prior existence of some language pattern is either necessary or sufficient to produce some thought pattern.

Brown, 1958: p. 260, via Pavlenko, 2014: 13

• This is a “radical transformation of Sapir’s and Whorf’s ideas into the SWH [Sapir-Whorf hypothesis]” ​(p. 13)
• “The change of paradigm was necessitated by the desire to make complex notions, articulated by linguistic anthropologists, fit experimental paradigms in psychology.” ​(p. 13)
• “First, they shifted the focus of the inquiry from the effects of obligatory grammatical categories, such as tense, to lexical domains, such as color […]. Secondly, they shifted from concepts as interpretive categories to cognitive processes, such as perception or memory, that were of little interest to Sapir and Whorf […]. Third, they privileged the idea of thought potential (and, by implication, what can be said) over Sapir’s and Whorf’s concerns with obligatory categories and habitual thought (and, by definition, with what is said). Fourth, they missed the insights about the illusory objectivity of one’s own language and replaced the interest in linguistic thought with independent ‘language’ and ‘cognition’. Last, they substituted Humboldt’s, Sapir’s and Whorf’s interest in multilingual awareness with a hypothesis articulated in monolingual terms.” ​(pp. 13 – 14)
• “[…] Brown and Lenneberg concealed the radical nature of their reformulation by giving Sapir and Whorf ‘credit’ for what should have been the Brown-Lenneberg hypothesis.” ​(p. 14)
• “Performing a feat reminiscent of Humpty Dumpty, Pinker (1994) made the SWH ‘mean’ what he wanted it to mean, namely ‘the idea that thought is the same thing as language’ (p. 57).” ​(p. 16)
• “the received belief in the validity of the terms of engagement articulated by Brown and Lenneberg and their attribution to Sapir and Whorf is still pervasive in many academic circles and evident in the numerous books and articles that regurgitate the SWH as the strong/weak dichotomy.” ​(pp. 16 – 17)

Bilingualism: Key Terms and Constructs §

• “Coordinate bilinguals, in this view, are speakers who learned their languages in distinct environments and have two conceptual systems associated with their two lexicons. Compound bilinguals learned their languages in a single environment and, consequently, have a single underlying and undifferentiated conceptual system linked to the two lexicons. Subordinate bilinguals, typically classroom learners who learned the second language via the means of the first, have a single system where the second-language lexicon is linked to conceptual representations through first-language words.” (based on Weinreich (1953) / Ervin and Osgood (1954)) ​(pp. 18 – 19)
• “bilingualism researchers define bilinguals as speakers who use two or more languages in their everyday lives, be it simultaneously (e.g., in bilingual families) or sequentially (e.g., in the context of immigration or study abroad).” ​(pp. 20 – 21)
• “To describe the order of language acquisition, researchers usually adopt a chronological approach and refer to the first (L1), second (L2), third (L3) language and so on […]. The term age of acquisition (AoA) refers to the age at which the L2 learning began, regardless of its context. […] The age of acquisition is not always the same as the age of arrival in the target-language (TL) context (also abbreviated as AoA or as AoAr). […] The term context of acquisition (CoA) refers to the context in which the language was learned, with a three-way distinction made between foreign language (FL) or instructed contexts, L2 or naturalistic contexts, and mixed contexts.” ​(p. 21)

The Evolution of Language §

The Evolution of Symbolic Cognition §

This amazing table ​(pp. 27 – 29) gives a timeline of the evolution of symbolic cognition in our species:

Pavlenko points out that this table reflects a Eurocentric bias because most intensively studied archeological sites are located in Europe. Recent discoveries in African sites show that “almost all cultural innovations documented in Europe are also present — at much earlier dates […]” ​(pp. 26 – 29). As new sites are discovered, the earliest confirmed dates for certain innovations keep getting pushed back.

The Development of the Prehistoric Mind §

• “Donald’s (1991, 2001) sequence begins with the episodic stage, where our Australopithecine ancestors displayed cognitive skills common for all primates: basic episodic and procedural memory and spatial representation. These skills enabled basic event perception, pattern recognition, situation analysis, imitation, and recall limited to the here and now. Their behavior was largely reactive to stimulus and the use of stones as tools opportunistic.” ​(p. 31)
• “the emergence of stone knapping marks the transition to the mimetic stage, where our ancestors communicated via gestures, vocal sounds, body language, and reenactments. The pressure to develop iconic signs that invoked the entities they referred to may have been created by the involvement in scavenging, which required communication of the location of food sources (i.e., dead carcasses) outside of the sensory range of message recipients and planning on how to get to them before other predators did (Bickerton, 2009). To be successful, mimesis would have involved conscious, intentional, self-initiated (auto-cued) embodied representational acts and would have required some self-awareness and the ability to control voluntary actions, regulate emotions, and imagine bodies in action (kinematic imagination).” ​(p. 31)
• “The next point at which we see rapid encephalization is the speciation of Homo sapiens and a gradual transition to what Donald (1991, 2001) calls the mythic stage, distinguished by narrative thought, complex linguistic skills, and the ability to create new symbolic systems. McBrearty and Brooks (2000) outline four types of changes that signal this transition and its accelerated pace of cultural development: (a) technological, where the replacement of the Acheulian industry by diverse blade technologies and hafted tools suggests new versatility, inventiveness, and capacity for logical thinking; (b) ecological, where successful expansion and colonization of previously uninhabited environments suggest planning depth and the ability to adapt and innovate; (c) economic and social, where intensification in resource use, including selective and tactical hunting and fishing, suggests greater planning depth, and the use of raw materials from far-away locations indicates scheduling, trading, and complex networks; and (d) symbolic, seen in adoption of symbolic objects in everyday life.” ​(p. 32)
• “The current theoretic stage is linked by Donald (1991, 2001) to the emergence of symbolic technologies of numeracy and literacy, which enabled paradigmatic thought and external memory storage, critical for development of theory and science. While the invention of writing is commonly dated to 3500 BC, its precursors are located much earlier in prehistory, when humans began using marks and notches on bones, stones, and wood for representational purposes (see also Chapter 3). Danziger (2008) argues that such ‘externally archived material is useful only to the extent that its organization is reflected in individual memory’ ​(p. 4). In this view, memory development involves the co-evolution of external memory devices, such as numbers or writing, shaped by the processing constraints of the human brain, and of new cognitive skills and functions, such as numeracy or literacy, necessitated by new mnemonic practices.” ​(pp. 33 – 34)

Language, Cognition, and Linguistic Thought §

• “If by ‘cognition’ we mean the processes of attention, perception and memory we share with other animals, we are undoubtedly justified in using ‘cognition’ as a discrete discursive construct, independent from ‘language’.” ​(p. 34)
• If […] one follows in the footsteps of Herder, Humboldt, Vygotsky, Sapir, and Whorf, one may arrive at a view of language as a form of cognition (Tomasello, 1999), a semiotic tool that co-evolved with cognition to facilitate everyday problem solving, complex reasoning, and communication in the mind that is embodied, extended, and distributed (Donald, 1991, 2001; Gamble, 2007; Renfrew, 2009). […] both context and embodiment play a central role in symbolic cognition and in the emergence of new symbolic systems.” ​(pp. 34 – 35)
• “In what follows, the notion of linguistic thought will refer to cognitive processes fully or partially mediated by language: categorical perception, numerical, temporal, and spatial cognition, segmentation and construal of events, narration, recall, and autobiographical memory, interpretive framing and self-mediation (inner speech), and affective processing and interpretation of emotional experience.” ​(p. 35)
• “What matters for the purposes of Whorfian inquiry is the way we use linguistic processes, and in particular the obligatory categories of our languages, to reach agreement on the interpretation of a temporarily shared social reality, the notion now known as intersubjectivity […].” ​(p. 35)

Linguistic Diversity and Language Change §

• “languages are inherently variable and that, in the context of synchronic variation, people favor certain options over others and sometimes simply innovate, leading to changes at all levels: phonological, morphological, syntactic, semantic, and pragmatic” ​(p. 37)
• “Both internal and external processes are explained by two types of factors, cognitive and social. Cognitive factors include (a) economy, i.e., the tendency to save effort, which may cause sound erosion, (b) analogy, i.e., the tendency to draw links between domains, which may give rise to new metaphors and eliminate paradigmatic alternation in favor of a single regular pattern, (c) expressiveness, i.e., the tendency to favor novel expressions in order to achieve greater effects, and (d) the need to reduce cognitive load, which, in bilingual speakers, may lead to regularization or convergence of two distinct categories or patterns (Silva-Corvalan, 2008). Social factors include but are not limited to discursive practices (e.g., vocabulary-related taboos necessitating lexical replacements), evolving social realities (e.g., technological changes leading to coinage and neologisms), and language ideologies and symbolic capital associated with particular languages or forms, which may lead speakers in language-contact situations to accentuate similarities (convergence) or deliberately foster differences (divergence) via new rules (e.g., flipping the gender of all masculine nouns to feminine and vice versa) (Evans, 2010).” ​(p. 38)
• “Geopolitically dominant languages spoken in contemporary industrialized societies, are learned by many L2 speakers — as a result, they tend to undergo simplification, e.g., encoding fewer grammaticalized deictic distinctions (Perkins, 1992). In contrast, languages spoken by small economically self-sufficient groups in relatively isolated areas are more likely to display greater morphosyntactic complexity and numerous exceptions learned by rote […].” ​(pp. 38 – 39)

Categorical Perception §

• “Grounded in basic cognitive abilities, such as memory and pattern recognition, categorical perception enables us to organize the ‘kaleidoscopic flux of impressions’ via rapid discrimination between stimuli, identification of individual stimuli as members of larger categories of objects, properties, relations, and phenomena, and activation of a large amount of information about these categories that, inter alia, allows us to judge typicality of individual members and instances (Brosch et al., 2010; Harnad, 1987; Rosch, 1977, 1978).” ​(p. 40)
• “The process of categorical perception is shaped by our ages and histories, our physical states (e.g., how tired or hungry we are), and our goals and needs (e.g., whether we are trying to accomplish a work-related task, find something to eat, or determine if it will rain).” ​(p. 40)

Referential Indeterminacy §

• “The studies to date show that the mapping of words onto external referents is a complex cognitive process that involves selective attention to particular features, functions, or properties of the object, event, or phenomenon to be named, matching of those properties with internal representations, and retrieval of words linked to these representations, which in turn involves selection between lexical alternatives […].” ​(p. 43)
• “even the most common objects may be named differently by different speakers (inter-speaker variation) or by the same speaker on different occasions (intraspeaker variation)” ​(p. 44)
• “Highly codable entities, in this view, have one standardized monolexemic (i.e., single-word) label used with high agreement (e.g., cup, green). In contrast, entities with low codability may elicit a variety of multiword descriptions and lexical alternatives (e.g., a kind of cup, chalice, color of sea grass, turquoise). Less standardized and lower-frequency terms, such as chalice or turquoise, may also be linked to somewhat different referents by different speakers. Both types of variation indicate weaker links between words and referents than in the case of highly codable entities.” ​(p. 44)

Color §

“Basic Color Terms” Theory §

• “[…] Berlin and Kay (1969) revitalized Geiger’s (1880) and Magnus’s (1880) idea that color lexicons evolve in a predictable sequence, from two terms, for dark and light, to eleven terms. Using experimental data from speakers of 20 languages and lexicographic data from another 78 languages, the researchers argued that languages do not divide the color spectrum in a random fashion – rather, the color spectrum itself is perceptually irregular, with focal colors more salient than non-focals. These focal colors serve as prototypes for basic color terms (BCTs), monolexemic or single-word items, psychologically salient to informants, whose meaning is not included in other terms and whose use is not restricted to a narrow class of objects.” ​(p. 47)
• “The BCT theory proposed that BCTs evolve around the same foci and in the same order.” ​(p. 47)
• However recent studies found “a great amount of variation in foci placement across and within language groups”, challenging “the idea of the perceptual salience of ‘universal foci’” ​(p. 48)
• “Most importantly, the critics of ‘color universals’ object to the ‘epistemological chauvinism’ reflected in the assumption that the abstract Western category of ‘color’ is ‘natural’ and in reliance on color terms of American English, which miraculously coincide with universal categories, and on the dimensions of hue, brightness and saturation, associated with the English terms (Lucy, 1992a, 1997a; Ratner, 1989; Saunders & van Brakel, 1997; Wierzbicka, 1990).” ​(p. 49)
• Also see: Wikipedia: Linguistic relativity and the color naming debate

(Lack of) Color Terms in Homeric Greek §

• Up until the 1930’s there were scholars believing that the ancient Greeks color vision wasn’t fully developed. ​(pp. 46 – 47)
• “Contrary to the Western view of ‘color’ as a natural property of the material world, many languages, including Bellonese (Polynesia), Mursi (Ethiopia), Pirahã (Brazil), Warlpiri (Australia) and Kalam and Yélî Dnye (New Guinea), do not encode color as an abstract dimension independent of other properties of material objects […] ‘Color’ as an abstract entity also has low communicative importance in these speech communities.” ​(pp. 49 – 50)
• “What is rendered invisible by the hue-based Western lens is the complexity of systems of visual description that combine references to hue with other locally salient distinctions, such as patterns, ripeness, texture, brightness, translucence, discreteness, humidity, shape, or location (Bricker, 1999; Bulmer, 1968; Conklin, 1955; Turton, 1980).” ​(p. 50)
• “Homeric Greek lacked a hue-based ‘color’ category (it did emerge in later centuries, as seen in Aristotle’s and Plato’s color theories) and relied on terms that emphasized the dark/light contrast, privileged luminosity, brightness, temporality, and movement, and were context dependent and imbued with ‘non-color’ meanings (e.g., khlôros [fresh, green foliage] describing honey or twigs as fresh, unripe, moist or full of sap) (Irwin, 1974; Lyons, 1995, 1999; Maxwell-Stuart, 1981). Translations that reduce Homer’s oinops [wine-looking] to the color reference, wine-dark sea, miss out on the other senses of the term that portray the sea as wine-eyed, drunk, rolling about, made choppy by strong winds, violent, dangerous, and ‘bloody’ in the sense of lethal (Maxwell-Stuart, 1981).” ​(p. 51)
• “Studies of other ancient languages similarly suggest that an abstract huebased category of ‘color’ evolved from systems that relied on analogies with natural dyes and concrete objects and privileged brightness and luminosity over hue; and in some cases, it was simply borrowed in the process of language contact (Biggam, 2007, 2012; Casson, 1997; Lyons, 1999; Warburton, 2007). Thus, the evolution of ‘color’ does not necessarily proceed from a simpler to a more complex system. Rather, as Woodworth (1910) suspected a long time ago, it may change from a more complex, elaborate, context-dependent, and variable system that requires speakers to pay simultaneous attention to a variety of features to a simpler abstract system that generalizes across a variety of entities, displays high naming agreement and relies on the combination of hue, brightness, and saturation, and sometimes just on hue.”

Language Effects on Perception of Color §

• People are better at remembering colors encoded in specific color terms in their language ​(p. 52)
• In a study participants had to push a button when they saw a square in between a stream of circles being shown to them. In one setup the squares were light green and the circles dark green — here native speakers of English and Greek reacted roughly the same. In an other setup the squares were light blue and the circles dark blue. In this setup, speakers of Greek — which has distinct terms for dark blue (μπλε / ble) and light blue (γαλάζιο / ghalazio) — reacted faster. “These results demonstrated that even in the context where Greek speakers did not have to discuss – or even pay attention to! – the colors, the obligatory linguistic distinction affected the early (pre-attentive) stages of perceptual integration and made the light/dark blue contrast more perceptually salient than the light/dark green one.” ​(p. 52)
• “Together, these findings suggest that color lexicons draw attention to the categories encoded in the language in question and make them more perceptually salient, facilitating color memory, learning, and categorical perception.” ​(p. 52)
• “The differences between English [with one basic term for blue] speakers on the one hand and Russian [with tow basic terms] and Ukrainian [with three basic terms] speakers on the other further suggested that speakers of languages with an obligatory light/dark blue distinction not only register this contrast more acutely, as demonstrated by Thierry and associates (2009), but also ‘ascribe’ more ‘significance’ to it in communicative contexts. In turn, English speakers – who undoubtedly can see and describe the distinctions in shades of blue – do not ‘ascribe’ the same significance to these distinctions and do not find them worthy of mentioning.” ​(p. 53)
• “Following Goldstone’s (1998) perceptual learning theory, the process by which speakers of ‘color’ languages develop the ability to focus on hue, to suppress irrelevant information, and to generalize across dissimilar entities, such as apples and emeralds, is viewed here as imprinting, whereby perception adapts to environment by developing specialized detectors (receptors) for particular stimuli. Studies with monolingual and bi- and multilingual speakers further suggest that acquisition of L1 color categories relies on the attention weighting mechanism (Goldstone, 1998), which increases selective attention paid to features relevant for everyday practice. In communicative contexts, this sensitivity translates into ‘ascribing significance’ to the categories in question, seen in their more frequent occurrence (Pavlenko, 2012a); it can also be identified in non-verbal tasks, such as the oddball detection task used by Thierry and associates (2009).” ​(pp. 63 – 64)

Bilinguals’ Color Categories §

• L1 influence on L2 naming patterns: “Ervin (1961) found that Navajo–English bilinguals favored the term yellow in the context where monolingual Navajos agreed on its translation equivalent Litso [yellow] and monolingual English speakers were split between yellow and brown. Bilinguals also displayed a preference for yellow in the context of the yellow/green boundary, where Navajo monolinguals used Litso [yellow] and English monolinguals favored green. […] Her findings suggested that the L1 influence on L2 may be particularly salient in contexts where the target language displays referential indeterminacy or variability in color naming patterns.” ​(p. 54)
• Internalization of L2 categories: Multiple studies showed that speakers can learn more fine-grained color distinctions made in a L2 ​(pp. 54 – 55)
• Co-existence of language-specific naming patterns: “[…] it appears that in some contexts bilinguals who use both languages on an everyday basis may be able to maintain distinct categories associated with their respective languages” ​(p. 60)
• L2 influence on L1 naming and categorization patterns: Color naming and categorization patterns from an L2 can impact the way color is talked about in L1, compared to monolingual L1 speakers: “Caskey-Sirmons and Hickerson (1977) found that bilinguals’ color foci and boundaries had shifted towards those of the L2 English.” ​(p. 61) Bilinguals may also stop using terms in their L1 that don’t exist in their L2 ​(pp. 61 – 62).
• “To date, only a few studies have examined factors that affect cognitive restructuring of bilinguals’ color categories. Studies of the maintenance of the L1 contrast (light/dark blue) in the context of the single L2 English category blue found that the degree of L2 influence on L1 was affected by the age of L2 acquisition (Andrews, 1994; Pavlenko, 2012a) and the length of stay in the L2 environment (Athanasopoulos, 2009; Athanasopoulos et al., 2010).” ​(pp. 62 – 63)

Literature §

• Pavlenko, A. (2014). The bilingual mind: And what it tells us about language and thought. Cambridge University Press.