Friday, August 30, 2013

A triumph for Whorf - an existential flip for the rest of us!


Imagine a world where the sky was green. Not because you were on a different planet but because you didn’t know the word ‘blue’… 

This may seem like the stuff of science fiction but empirical evidence over the years is now suggesting that our words might indeed be shaping our thoughts and perceptions. 

Benjamin Lee Whorf, a linguist and a fire prevention engineer, was one of the pioneers to put forward this idea of linguistic relativity with Edward Sapir - his mentor at Yale.  Their fundamental idea was that linguistic categories influence perception and cognition or putting it simply - our language and our words influence our thoughts and our perceptions. 
Yes. You read it right. 
Our words influence our thoughts and our perceptions. Not just the other way around.

Sapir and Whorf's idea of linguistic relativity was explicitly drawn on Einstein's principle of general relativity such that the grammatical and semantic categories of a specific language provide a frame of reference through which the observations are made. They were deeply influenced by the ideas of Bertrand Russell and Ludwig Wittgenstein, whose view was that natural language potentially obscures, rather than facilitates, the mind's ability to perceive and describe the world as it really is. In the mid-twentieth century, this argument built the case for the extensive use of formal logic, to question the very fundamentals of logic and mathematics and to arrive at a set of axioms based on rigorous thought and didactic reasoning. Whorf  built on the ideas of Nietzsche and Wittgenstein and developed the theory of linguistic relativity but it was based on little empirical evidence and thus fell out of with the intellectual community within a decade of his death. In one oft quoted passage, Whorf writes: 

"We dissect nature along lines laid down by our native language. The categories and types that we isolate from the world of phenomena we do not find there because they stare every observer in the face; on the contrary, the world is presented in a kaleidoscope flux of impressions which has to be organized by our minds—and this means largely by the linguistic systems of our minds. We cut nature up, organize it into concepts, and ascribe significances as we do, largely because we are parties to an agreement to organize it in this way—an agreement that holds throughout our speech community and is codified in the patterns of our language. The agreement is of course, an implicit and unstated one, but its terms are absolutely obligatory; we cannot talk at all except by subscribing to the organization and classification of data that the agreement decrees. We are thus introduced to a new principle of relativity, which holds that all observers are not led by the same physical evidence to the same picture of the universe, unless their linguistic backgrounds are similar, or can in some way be calibrated."

But can this really be true? Can our language determine our perception of the universe?
Apparently, it can. And, here is some of the evidence for it that has compelled me to make that existential flip in my thinking and perception.

1) Blue or Green - the brain sees it only when it 'knows' it 
In a classical debate, cognitive scientists have been trying to establish the relation between language and perception. At one end of this debate is the strong -form Whorfian-theory that says that our perception of the world is shaped by the semantic categories of our native language and that these categories vary widely across languages. At the other end of the debate spectrum, is the universalist stance which claims instead that there is a universal repertoire of human thought and perception which leaves its imprint on all the languages of the world. Over the years, evidence has swung back and forth but empirical evidence from the domain of color and color-perception were some of the first and most comprehensive to weigh in on this rather fundamental question.

The effect of language on thought and perception was first tested with color perception in a 2006 study by Gilbert et al. Their study attempted to probe the perceptual discrimination of colors that straddled the boundary of blue and green - a boundary that exists in the English language but is absent in many others. The question was: does having a label for a color made a difference in how we see it?

As part of the experiment, English (American) speaking participants were presented with a central fixation cross with a ring of colored squares tiled around it. All the squares in the ring were of the same color except for the target. The participants were required to identify the differently colored square and indicate its location on the left vs. right side of the ring by pressing a button with the corresponding hand. The test was designed in such a way that the target color had either the same name as the color of the other squares (eg. green against the background of a different green) or a different name (eg. green against a background of blue) i.e. belonged to the same category or to a different category. 


Termed as Categorical perception (CP), this test was designed to see if having a different word/name for a color improved our ability to perceive it especially when the stimuli lie at the edge of a boundary. 

Instead of a simple yes or no answer the experiment yielded a rather interesting pattern: The subjects were better able to identify cross-category targets faster than same-category targets but only when presented in their right visual field, thus suggesting that Whorf's hypothesis might have been right after all; atleast half of it. 

Interestingly our language faculties are localized in our left hemisphere and the brain functions contra-laterally (left side of the brain controlling the right visual field). To test the effect of language, a parallel task requiring the use of verbal resources was added to the test and interestingly the effect on categorical perception by the left hemisphere (in the right visual field/RVF) was lost and in fact, was reversed. They also found that such a reversal was not seen when a non-verbal task of similar difficulty was added, thus suggesting that this lateralization effect was because of our language indeed. 


As one would expect, split brain patients whose left and right hemispheres are unable to talk to each other lost any trace of CP bias in the left visual field/LVF.

Interestingly at the time, a separate study by the same group expanded this paradigm to the use of non-color stimuli, namely silhouettes of dogs and cats thus extending the effect of our words on the way we think and perceive the world.  

But then what happens in children who are as yet uncorrupted by language? Do children who cannot yet talk, see and think differently from children who can? 

This is a rather interesting question and was addressed in a study by Franklin et al in 2008 as they first compared infant and adult performance on a visual search task much like that used earlier by Gilbert's 2006 study. As seen previously, adults showed dominant CP in their right visual field influenced by the language center in the left hemisphere. The pre-linguistic infants however showed no such CP bias in the Right visual field (RVF) and instead exhibited a clear CP in the left visual field (LVF). Thus, it seemed that as the children grew up and acquired language, there was a migration in their perception of categories from the RH/LVF (as infants) to the LH/RVF as adults (with language). 

Instead of comparing two very different groups - the infants and the adults, Franklin et al, next compared toddlers (2-5 yr old) in two groups: the learners and the namers. While the namers already had acquired their color terms, the learners were still finding their way around the color spectrum. Interestingly, the learners patterned colors like the infants while the namers showed CP like the adults. This strongly suggested that learning the color terms causes a shift in our categorization of color perception from the right hemisphere to the left. 


What seemed like green to start with suddenly starts looking different once we learn the word blue.  

When investigated further, these differences in color perception were further mimicked at the deeper level of electrical activity in the brain. Functional MRI studies have found that discriminating colors of different lexical categories (vs. the same category) elicited a faster and stronger response in the language regions of the left hemisphere especially when the colors were presented in the right visual field. Based on these studies over the past decade, it thus seems uncontroversial now, that once language is learned, the language labels replace any existing categories and shape our perceptual discrimination (especially in the left hemisphere/ right visual field).

2) Mapping our space-time paradigm - one word at a time

For a long time in the past, the impact of language on human thought has remained an intriguing question in the realm of psychology, philosophy and linguistics. Very little was done to empirically test any of these claims. However, over the past decade or so, researchers like Lera Boroditsky - a psychologist and neuroscientist at Stanford, have dared to jump in and get their hands dirty in trying to arrive at more definitive answers. Research in Dr. Borodistky's labs at Stanford and MIT is focused on trying to test the hypothesis that language does shape our thoughts, above and beyond our abilities to see color or shape. They have collected data from around the world: from China, Greece, Chile, Indonesia, Russia and Aboriginal Australia and have found that people who speak different languages do indeed think differently. 

Given some thought, this effect of language is rather evident because, after all, different languages tend to provide different levels of information - gender of the subject/object, tense, verb information etc. While some languages like Russian alter the verb to indicate the tense and the gender, some others like Turkish would also indicate how the information was acquired. Some other Romance languages like Spanish and french, also ascribe gender to common nouns - thus making a table masculine or feminine. In fact, some Australian Aboriginal languages have upto sixteen genders that include classes of hunting weapons, canines, things that are shiny, etc. 

Since languages communicate different information in different forms, it is also conceivable that the speakers might end up attending to, partitioning and remembering their experiences differently by virtue of their different languages? Questions like these have been addressed by the work done by Dr. Lera's group and others and have brought us farther along this line of questioning. 

Consider the following examples.

In northern Australia lies Pormpuraaw - a small aboriginal community on the western edge of Cape York, where the locals - Kuuk Thaayorre, have a rather interesting approach to space. Unlike us, english speakers, who define space relative to an observer by the use of words like right, left, front, back etc, the Kuuk Thayorre, like many other Aboriginal groups use the cardinal direction terms - North, South, East and West - to define space. All the time. All space in this community - from a village to an arm or a leg is specified along cardinal directions. Not only does it save them from that moment of confusion we all have, when someone facing us asks us to look right or left, and we wonder - "my left or your left"; this practice also makes them rather capable navigators. This constant training of their attention to the geographical coordinates forces them to pay exquisite attention to their geographical coordinates - all the time. 

But how does the use of cardinal coordinates really affect thought - not just navigational ability? And this is where Boroditsky and her group design some simple experiments to demonstrate that since space is such a fundamental domain of thought, differences in spatial perception also tend to impinge on our other more complex and abstract representations too. Surprising as it may seem, it appears that our representations of things as time, number, musical pitch, kinship relations, morality, and emotions depend on how we think about space. For instance, if people are given pictures that are snapshots of a temporal progression (eg. a man aging or a fruit being eaten) and asked to arrange them chronologically - one can see some very interesting trends. English speakers arranged the cards such that time proceeds from left to right. Hebrew speakers however arranged the cards from right to left (their script progresses from right to left). The Kuuk Thaayorre on the other hand, lacking such concepts as left and right followed a completely different trend. Their arrangements were not random; instead they arranged time from east to west (sunrise to sunset, perhaps?). So, when facing south, they arranged the cards from left to right; and when facing north, they went from right to left, and so on.  

People's ideas and representations of time also differed in other ways such that, while english speakers tend to talk about time using horizontal spatial metaphors (ahead, behind etc), mandarin speakers use a vertical arrangement for time, in keeping with their script. When subjects were given a spot for today and asked to plot yesterday or tomorrow, english speakers nearly always pointed horizontally while mandarin speakers pointed on the vertical axes. In fact, Chinese calendars move downwards and across (right to left on the page) like the written script itself.

In addition, research also shows a strong effect of language on some of the basic aspects of time perception. For example, since english speakers prefer to talk about time or duration in terms of length (short talk vs. long) their perceptions of time tend to be confused by distance information, such that they estimate a longer line to have remained on the screen for a longer period of time, although the two parameters are completely unrelated. 

But these tests are plagued by one critical question - the question of causality. How do we know that these people of different cultures are perceiving things differently only because of their language and not because of anything else. Reason would dictate that if these people learnt a new language, their perception of the world would change accordingly. And so, in one such study, English speakers were taught to use size metaphors like in Greek to describe duration or vertical metaphors like in Mandarin to describe event order. Remarkably, once the english speakers had learnt to talk about time in these new ways, their cognitive performance began to resemble that of native Greeks and Chinese thus establishing a rather conclusive role for language in instructing how we think.  

3) Word games - when the same 'key' can become heavy or tiny! 
The sub-conscious impact of language also extends from the realm of abstract concepts like time and space into the more physical. 

A clear case in point is that of grammatical gender as seen in romance languages like Spanish, German and French where nouns are assigned masculine or feminine genders. Speakers, in turn have to change pronouns, adjectives, verb endings, possessives, numerals and so on, depending on the noun's gender. 

But do these subconscious perceptions subsequently bias our opinions of everyday objects? When tested, it was indeed found to be so. In one study from Boroditsky's group, the experimenters asked German and Spanish speakers to describe objects having opposite gender assignment in the two languages to see how the gender influences perception. When asked to describe a "key" - a word that is masculine in German and feminine in spanish - the German speakers used words like "hard", "heavy", "jagged", "metal", "serrated", "useful" etc while the spanish speakers were more likely to use words like "golden", "intricate", "little", "lovely", "shiny", and "tiny". 

Interestingly, these results emerged even though the testing was done in a third neutral language like english. The same pattern of results also emerged in entirely non-linguistic tasks like when subjects had to put images together. 

Apparently, even small flukes of grammar, like the seemingly arbitrary assignment of gender to a noun, cal dramatically alter our perception of the world. 

4) Honest lies - true testimonies of events that never happened
Honest lies are factually inaccurate information that we have told with all honesty because that's what we remember them to be. We have all told them inadvertently because of when, how and where were asked the question. 

Innocent people have been caught on the wrong side of law while liable culprits and criminals have sometimes been set free because of this strange malleability of the human mind. And no one paid much attention to study this phenomenon because we were all oblivious to how our mind was being tricked and how it was in turn tricking us, until of course, a psychologist from UC Irvine, Elizabeth Loftus dived in and highlighted some of these hard-to-swallow facts. 

Loftus' journey began at a social-psychology class when she saw that people could name a 'yellow bird' faster than a 'bird that's yellow'. Her subsequent search for interesting puzzles, funds and opportunity, brought her to the US department of transportation where Loftus began her research into car accidents and the limitations of eye-witness testimony. She empirically showed that people's eye-witness testimonies and versions of accidents can vary depending on the wording of the questions asked of them. She showed people clips of car accidents and asked them to estimate the speed of cars. People when asked, "How fast were the cars going when they smashed into each other?' gave higher estimates on average than those with whom the verb "hit" was used. Those who heard the verb 'contacted' in the question gave the lowest estimates of the speeds again confirming the idea that words and language can mould and shape our thoughts. 

The effect of words and language on our thinking does not merely extend to how we judge unknowns like the speed in this instance - they can also mislead us into fabricating parts of the story. Those asked about cars smashing into one another were more than twice as likely as others to report seeing broken glass when asked about the accident a week later, even though there was none in the video. This clearly suggested that our words and language can sub-consciously lead us into confabulating 'facts' and contaminating our memory traces. 

After these dramatic beginnings, Loftus went on to publish several other studies that showed how memories can be contorted and eye-witness accounts can be tainted. She has gradually extended her work from the realm of pure academic interest into having social ramifications and is currently pushing for broader legal reforms. While the focus of Loftus' studies is on the unreliability of our memories, her work is also relevant to the effect of language on thinking. As can be seen from her experiments, the exact phrasing of the question had a profound effect on what and how people thought and recalled events. 

5) Gender personification, abstraction and art 
This is a rather interesting idea that explores the effect of native languages on an artists perception of abstract concepts in art such as death, victory, sin, time etc. When representing such emotions in human forms, how does an artist decide the gender? Accordingly to Dr. Boroditsky, it appears that in most such abstract personifications (85% apparently), whether a male or a female figure is chosen is predicted by the grammatical gender of the word in the artist's native language. Not too surprising, perhaps, once we make that Whorfian existential flip. 
6) To see or not to see - It's all in the dictionary! 
The many examples so far establish a critical effect of words on our thinking, on our ability to categorize and discriminate. But, the effect of language on perception itself remains. To what extent is the awareness of an object affected by factors outside of our vision? The traditional view is that although a number of non-visual factors can affect where and what one attends to, the very content of what we see is not susceptible to outside influences - but is determined by physical factors. 

However, a recent study published in the Proceedings of the National Academy of Sciences, does a lot to overthrow this paradigm. Through empirical evidence, the authors demonstrate a strong effect of language on perception itself; an effect divorced from any higher level cognitive processing that one might suspect. They show that auditory linguistic labels can not just affect what one sees but whether one sees something at all. Their studies show that an otherwise invisible kangaroo can be boosted into awareness by language, by using the word for a kangaroo while seeing the stimulus. 

This idea that a "higher-level" process like word recognition or language can influence a "lower-level" process such as visual perception presents a challenge to our normal feed-forward modeling of cognition and perception; where it is assumed that the world is seen for what it is but then our brains filter out and attend to specific stimuli. 

Although many studies have demonstrated a clear effect of language on post-perception decision making, an outstanding question in the field is to discriminate the effect of language on perception vs. perceptual identification. One way to do this is to keep visual stimulus (object) unchanging while manipulating the top-down influences that change our perception of the object.  This way, any differences in perception by stimulus-driven processes can be ruled out. 

In this study, the authors investigated whether hearing a language label can affect our ability to simply detect the presence of an object. In order to suppress conscious visual awareness along with semantic processing, the authors use a process called as interocular rivalry i.e. they present different stimuli to the two eyes, such that the stimuli compete and the same retinal input can give rise to different conscious percepts. Which of the two stimuli gain our attention is determined by visual aspects of the stimulus - contrast, luminosity, size, eccentricity, contour density etc.

Continuous flash suppression (CFS), a variant of interocular rivalry, is particularly useful to testing visual awareness and forms the basis of these experiments. In continuous flash suppression, an object is placed in interocular competition with high-contrast noise patterns (alternating really fast at 10 Hz) thus suppressing the real stimulus from awareness for extended periods of time. 
The authors began by testing if hearing a verbal cue or language label can make an otherwise invisible object (invisible by CFS) to pop-up and become visible? They reasoned that if words can change our perception of things, then hearing a label before seeing a picture (that is normally invisible due to CFS), should increase the likelihood of seeing the picture. And indeed, hearing a word label (like a kangaroo) before the simple detection task increased performance or increased the ability to detect the "kangaroo". Further, valid labels improved the detection while invalid labels actually decreased performance. As can be seen in the graphs below, these word labels affected both the sensitivity and the speed of the detection. Also, interestingly the study showed that the effectiveness of a particular label varied predictably as a function of the match between the shape of the stimulus and the shape denoted by the label. 


Unlike the previous studies, this report clearly shows that language affects the very visibility of objects. The labels provided by language affect performance not just on tasks requiring processing (explicit identification, categorization or discrimination) of what we see but can actually make us see something that did not exist before.

While one interpretation would categorize such a perceptual bias as entirely maladaptive from an evolutionary standpoint, the alternate possibility suggests that such top-down fine-tuning of perception can make it more sensitive to stimuli that are relevant to what is needed. Such a permeable perceptual system would allow for influences outside of vision (such as language) to exert an influence and to modify both perception and behavior; such as looking for a lion, when someone screams lion behind you.    

This study clearly demonstrates that as was suspected by Whorf, our languages keenly influence not just what we think but also what we see. 

Linguistic relativity - started off as an obscure theory, unsupported by facts and was soon a relic in the textbooks. But empirical evidence from half a century later has not just resuscitated the theory but has re-established it as a cornerstone in the field. 

It seems now that Wittgenstein was well ahead of the rest of us when he said, "Philosophy, is a battle against the bewitchment of our intelligence by means of language." In time however, we should learn enough about these spells to free our minds and our thoughts. 

References: 

    Corrupted memory, Nature, 14 August 2013, A feature on Elizabeth Loftus by Moheb Costandi
     Regier, T., & Kay, P. (2009). Language, thought, and color: Whorf was half right; Trends in Cognitive Sciences, 13 (10), 439-446
    Gilbert AL, Regier T, Kay P, & Ivry RB (2006). Whorf hypothesis is supported in the right visual field but not the left. Proceedings of the National Academy of Sciences of the United States of America, 103 (2), 489-94
    Franklin A, Drivonikou GV, Clifford A, Kay P, Regier T, & Davies IR (2008). Lateralization of categorical perception of color changes with color term acquisition. Proceedings of the National Academy of Sciences of the United States of America, 105 (47), 18221-5
    Radiolab, a public radio program from the WNYC did a wonderful one hour pod-cast on color perception that you might want to listen to
    Wikipedia entry on Whorf.
    Language can boost otherwise unseen objects into visual awareness, Gary Lupyan and Emily J Ward, Proceedings of the National Academy of Sciences, June 25, 2013 
    Diamond, J (2010). Science, 330 (332) The benefits of multilingualism
      A more abridged version of this post was published in Nature India and here is the link for it. 
      Although, I like the crisper version too and appreciate the edits, I like the depth and flow of this one.