To Notice and to Learn

Observations on ideas, human mind, and the world around us

Summary: The Language Instinct, by Steven Pinker

There has been a revolution in the field of linguistics over the past 50 years, initiated by the work of Noam Chomsky.  Prior to him, social scientists thought, consistent with the dominant Behaviorist perspective of the day (and also with common sense), that children picked up language from those around them through simple exposure, imitation, and reinforcement.  Chomsky was the first to question this, and raised as counter-evidence the rather obvious observation that as children start using language they often utter things that they almost certainly did not hear, such as “I eated the spaghetti yesterday” and “Why are you joking me so much?”  It thus appeared that they were attempting to apply rules rather than imitating what they heard.  This insight has led to a re-conceptualization of the field of linguistics and there has been an explosion of research on how grammatical rules are learned and what aspects of language might be universal across different cultures.  There are many hypotheses, unsettled questions, and controversies in the field, but there have also been many insights, and I think we in the general public have not had enough exposure to these insights.

The most noted popularizer of the findings of modern linguistics is the best-selling author Steven Pinker, of Harvard University, who I only stumbled upon a couple of years ago (at that cathedral for book-lovers, Powell’s Bookstore in Portland, Oregon).  Pinker has written two fantastic books that I want to discuss in this post:

The Language Instinct: How the Mind Creates Language” (1994) (The Language Instinct on
“The Stuff of Thought: Language as a Window into Human Nature” (2007) (The Stuff of Thought on

These books shed light on how our minds work with respect to this incredible and unique ability we have to utilize advanced language to communicate with each other.

In this post, I will offer my summary of the points that I found most compelling.  First, I want to list four things that modern linguistics tells us about language (Part 1, this blog post), and then I’ll describe three areas where language sheds light on how we think (Part 2, a future blog post):

1.  Language is something most of us take for granted and is much more complex than we realize.  If you take a step back and picture two humans standing next to each other, hardly moving, yet through speaking and listening able to transfer what is in one mind to the other and to affect each others’ thoughts, via only language, it is quite remarkable.  Here is the sequence of events:  Person A thinks of something they want to communicate, chooses the vocabulary and syntax of their shared language that they want to use, and then uses the air and cavities of his throat, tongue, oral cavity, nasal cavity, teeth and lips (each of which have other primary functions) to make certain sounds in a specific sequence (a digital to analog conversion).  Person B hears this sequence with her ears, and decodes the noises into words (an analog to digital conversion), analyzes the grammar, and extracts the meaning.  This process is so robust that they may be doing this apparently effortlessly, while constantly interrupting each other, and often finalizing their thoughts only when they are in the midst of uttering their sentences–all while communicating very effectively.  The process is so flexible with respect to content that two people sharing a language could be talking about what it’s like to grow up with an emotionally unavailable parent, justification for their cheese and wine pairing preferences, or mathematical equations describing the few initial seconds after the Big Bang.  This showcases the dazzling capabilities of human language.

One of the important features of language is a flexible combinatorial grammar.  To illustrate this, let’s take as an example this fairly straightforward sentence in English:  “The man with the monstrously ugly umbrella left the house“.   Below is how it would be diagrammed by most linguists to analyze the sentence (note that this is not necessarily the same as the diagramming trees that English teachers use to teach ‘proper’ grammar):



The labels here are: S = sentence, NP = noun phrase, VP = verb phrase, PP = prepositional phrase, D = determiner, etc.  When the sentence is shown in this way, we see the complexity of the structure required to both construct and decode this sentence—yet it is this structure that allows us to easily modify it in an unlimited number of ways.  For example, if we wanted to characterize the house further, we could add structure to the right-most noun phrase (NP) and say “the old, yet well-kept house”.  Or if we wanted to describe the man further, we could add a whole new PP phrase on the left and say “man from the historic town of Alexandria” or “man whom I met yesterday”.  we could also wrap this whole sentence into a phrase and start the sentence with “I’m thinking of writing a novel about the man…”  The possibilities are limitless, and this is why we can comfortably hear or read a sentence that is 80 or 100 words long and still make sense of it with only minimal conscious effort.  We often under-appreciate the complexity of grammar, because, Pinker says:

“Ordinary speech, like color vision or walking, is a paradigm of engineering excellence—a technology that works so well that the user takes its outcome for granted, unaware of the complicated machinery hidden behind the panels.” (Pinker, 1994, p.15)

Also, in other languages besides English there are many common aspects of how their grammars work, such as the concept of modifiers, embedded phrases, etc., and there are theories of so-called “Universal Grammar” that apply to all (or nearly all) known human languages.

There is another aspect of grammar that is impressive: long-distance dependencies, when, in English, we use if-then, either-or, or use case agreement (singular, plural) across words that are separated by long phrases.  Here is a wonderful example that Pinker cites from TV Guide and analyzes:

“How Ann Salisbury can claim that Pam Dawber’s anger at not receiving her fair share of acclaim for Mork and Mindy’s success derives from a fragile ego escapes me”

“At the point just after the word ‘not’, the letter-writer had to keep four grammatical commitments in mind: (1) ‘not’ requires ‘-ing’ (her anger at not receiving acclaim); (2) ‘at’ requires some kind of noun or gerund (her anger at not receiving acclaim); (3) the singular subject ‘Pam Dawler’s anger’ requires the verb fourteen words downstream to agree with it in number (Dawber’s anger…derives from); (4) the singular subject beginning with ‘How’ requires the verb twenty-seven words downstream to agree with it in number (How…escapes me).  Similarly, a reader must keep these dependencies in mind while interpreting the sentence.” (Pinker, 1994, p.89)

Linguistic analysis also sheds light on many apparent quirks of language, and usually shows that there is some rationale behind them.  Take for example, these two words: ‘mouse-infested’ and ‘rat-infested’.  We know the plural of ‘mouse’ is ‘mice’ and the plural of ‘rat’ is ‘rats’.  But when we make the first two words plural, we would likely say ‘mice-infested’ and ‘rat-infested’ (not ‘rats-infested’).  Why is this?  Well, apparently, our brain does a look-up of words in a ‘mental dictionary’, and the entries store stems and irregular plural forms, but does not store the regular plural; instead an ‘add –s’ rule is applied on the fly when needed.  When we make compound words, we take the stem words and join them before the regular plural rule is applied—but in the case of irregular plurals, we do pick them up in the mental dictionary entry, and thus ‘mice’ gets used in the compound formation.  Fascinating!  This is also why the plural of ‘toothmark’ is ‘teethmarks’ but the plural of ‘clawmark’ is not ‘clawsmarks’, because ‘tooth’ has an irregular plural but ‘claw’ does not.  This insight reveals an intricate picture of how our minds process language, along with the more surface thrill of explaining an interesting linguistic quirk.

2.  We acquire language effortlessly and may even have a specific instinct for it.  There are clearly some developmental periods where learning a native language is easy, and nearly all humans, regardless of IQ, socio-economic status, or parental rearing philosophy pick up their native language fluently.  At the age of around three, children start speaking in fluid complex sentences.  If a foreign language is learned before the age of about 12-13, the person can learn to speak it without an accent, whereas if a foreign language is learned after that, there will nearly always be an accent (an instance is Arnold Schwarzenegger, who has been in the U.S. and speaking English for 45 years).    There is also experimental research that shows that 6-month old infants can audibly distinguish different sounds across the worlds languages, but after 6-months lose the ability to distinguish those sounds that are not in their native language (See the fascinating TED Talk on the linguistic genius of babies).

Pinker makes the case that humans have a specific ‘instinct’ for acquiring and using language, and supports it with three areas of evidence.  First is the poverty of the input argument which states that children do not experience sufficient input from their environment to develop the complex rules and structures of language—they are exposed to various words and sentences, but then must generalize to the complex rules of grammar, and they generate new creative sentences that generally follow these rules, well before they go to school and learn to diagram sentences.  The second area of evidence is creolization, which is what happens when a pidgin language (a rough, simple patchwork of communicative phrases used by different linguistic peoples who are thrown together by some historical circumstance) is learned by children—they are turned into a full-blown language with complex rule-based grammars, and the pidgin develops into a creole.  This shows evidence of a native linguistic capability, and as Pinker puts it:

 “…complex language is universal because children actually reinvent it, generation after generation—not because they are taught, not because they are generally smart, not because it is useful to them, but because they just can’t help it.” (Pinker, 1994, p.20). 

The third area of evidence for the language instinct are language impairments that are due to injury (aphasia) or developmental disabilities (specific learning impairments) where individuals may be unable to use or comprehend certain types of grammar, but seem otherwise to have normal intelligence.

I think that Pinker presents a compelling case for a language instinct, particularly when you reflect that nearly everyone, in a wide variety of environments, develops a mastery of their native language effortlessly, and can discern fine points of grammar and syntax even without fully realizing it (and without knowledge of syntactic trees, etc.)  I was always struck by the advice, when learning a foreign language, to “check with a native speaker” to see if something is correct—not a language instructor, or even someone used to explaining how language works, but any random person walking the street who happens to speak that language!  This points to what an accomplishment shared languages are, and I would not be surprised if was confirmed that there is a specific, dedicated biological basis for it.  I should mention, however, that the idea of a biological language instinct is controversial in the field of linguistics, and some linguists advocate the view that we acquire language through a general-purpose learning mechanism powered by a flexible intelligence.  In either case, however, our ability to learn language is clearly remarkable, and ironically it is the very ease with which we wield it which leads us to take it for granted.

3.  Despite alarmists who think language use is degenerating among the general public (and especially youth), everyday language usage is usually sophisticated and rule-governed.  Let’s start with the big picture here:  the purpose of language is to communicate across minds, so if this is happening when people use language, and people able to communicate complex thoughts to one another, their use of language is doing its job.  Now it may be the case, such as with slang, that some sub-groups adopt conventions that other groups are not familiar with, and thus parts of their language may seem unintelligible (I would bet that the same happens when every generation hears the style of music popular with the next generation).  But upon rational analysis, the conventions of slang have definite meanings and rules of usage.  Believe it or not, this even applies to texting (here is a TED Talk on how a linguist is looking at texting, by John McWhorter–he’s excellent as I’ve also watched his overview of linguistics lecture series, produced by The Teaching Company).

There is a distinction between descriptive grammar and prescriptive grammar, and linguists focus on the former, which is how language is actually used by people (which is wonderfully complex and effective), and consider the latter to be of minor importance, largely consisting of overly strict rules (regarding the splitting infinitives, ending in a prepositions, etc.) that are imagined to lead to greater clarity (for people who were not confused in the first place).  It is a fact that language evolves over time through usage, adapting to new contexts and shedding unused forms, and nearly the whole time there are people who fret over any of these changes and would prefer to embalm a language in its present pristine form forever.  But this is a myopic view; there are many ways that English has changed that were undoubtedly bemoaned by people of the time.  For example, the Early Modern English second-person singular/plural pronoun distinction of ‘thou’ and ‘you’ has been collapsed into just ‘you’, and with it we lost the ability to make the distinction of referring to one or more than one person in the second-person (which is distinguished in many languages, as well as in the “y’all” of the American South).  But in dropping ‘thou’ we also were able to jettison the related words ‘thee’, ‘thy’, and ‘thine’, and thus benefit from greater simplicity.  It is also hypothesized that our past tense marker ‘ed’ might have been a collapse of a word combination adding ‘-did’, and thus ‘hammer-did’ turned into ‘hammered’.  The point is that a language never stays still, and is ultimately responsive to the needs of the people who use it.  There does appear to be a continual movement towards making things easier for speakers (historically it evolved towards the commoners usage patterns), but the sophistication of a language is safeguarded by the fact that it’s users will demand from it the ability articulate their most complex thoughts, and thus a language will always contain intricate, nuanced, and subtle features.

Another observation from the viewpoint of linguistics is that dialects are full-fledged languages with regular rules of grammar and phonology that can be cataloged.  I think we sometimes feel that if a language variant is derived from another, it is somehow corrupted, or at least inferior in some way to the source language, even though all of our modern languages are derivatives from some precursor language.  When linguists study dialects, they see the full panoply of structures and rules, and invariably, there are always some patterns in the dialects that are more linguistically complex than those in the source language.  This is the case, for example, in the dialect known as Black English Vernacular (also known as African American Vernacular English), where you can have two statements:  “He be workin’.” and “He workin’.”  The first denotes a current action, but the second denotes a habitual, i.e. a steady job, which he may or may not be doing at this instant.  In Standard English, the same phrase is used for both “He is working”, and does not distinguish between the two meanings.  Thus, linguistics provides an impartial viewpoint that can help dispel our biased view of dialects.  I am reminded here of the coining of the term ‘barbarian’ in Ancient Greek, which referred to, ahem–all non-Greek speaking peoples!

4.  Language is not the same as thought and we are not completely ‘confined’ by our native language.  I remember in college learning about the Sapir-Whorf hypothesis that a people’s language drastically affects how they think, with the examples of the number of words Eskimos have for snow and the Hopi’s lack of words for time (both have been ‘debunked’ as fanciful exaggerations).  The conclusion that language determines thought is quite congenial to the standard social science paradigm because in that case people are entirely the products of their culture, and culture can be criticized/reformed without laying the blame on individuals.  But of course, hoping something is true does not make it so.  Pinker marshals several strands of evidence that indicate that thought is not the same as language:

  • Infants who have not yet developed language skills have been shown, through experiments, to have thoughts regarding cause and effect, simple counting, and the conservation of matter (e.g. in pouring water from a tall to a wide glass)
  • People also think in images, as when we compare the shapes of objects by mentally rotating them in our minds.
  • New words are created (neologisms), when existing words in a language aren’t up to the job
  • We are able to create high fidelity translations from one language to another (excepting poetry, or evocative prose, perhaps)
  • We sometimes struggle to express our thoughts and to “find the right words” that match our thoughts
  • When ‘euphemisms’ are pointed out to us, we are not such prisoners of the words that we don’t see through them (e.g. ‘headcount rationalization’, ‘revenue enhancements’,  ‘opportunity for development’, ‘collateral damage’, ‘I like you as a friend’, etc.).

I think this is compelling evidence to show that thought is not confined by one’s language (though can of course be influenced by it), and I think there are two upsides to this:  first, it means that there may be some commonalities in the way we think as humans, which is an important area worth studying (Part 2 of this post will consider some findings from Pinker’s The Stuff of Thought), and second, we don’t have to view language as a prison that confines us, but can continually strive to find better ways to express our thoughts, which also usually sharpens our thinking on the topic.

One type of observation on comparing languages has always puzzles me: when people point out that “there’s no word in language X that is equivalent to it in language Y—it’s untranslatable”.  Now I, along with everyone else, finds these cases to be interesting and insightful, but why do we set the bar so high that we expect a one-word for one-word translation of every concept?  Maybe the more reasonable question to ask is: Can we translate the concept accurately using a short sequence of words?   I would assume that 80% of these cases would instantly become much less interesting.  Even in cases where there are various meanings, you could likely pick a translation that is appropriate for that context, much as translators do.  Don’t get me wrong, I don’t want to deny human differences, I value them and want to learn from them, etc.; I just happen to think that we have much more in common with each other than we have differences, but that we sometimes are overly fixated on the differences, become enamored with the ‘exotic’, and create more separation between ourselves and other groups of people than is warranted.

If you want to hear Steven Pinker give a brief summary of The Language Instinct, here is a video taped in September, 2011.

In Part 2 of this blog post, I will summarize some points from The Stuff of Thought, which I found even more interesting than the Language Instinct—Pinker points out areas where language gives us clues to how our minds conceptualize the world of space, time, and people in the world around us.



Categories: The Human Mind and Psychology

Tags: , , , , ,

7 replies

  1. Hi Charlie,
    Thanks for dropping by and the follow.

    As for your question about my thougts on Language Instinct, well I come the “enemy camp” on this one. Cognitive Linguistics (CL) grew out out of a reaction against it and has proposed the embodied mind, usage and conceptual metaphor asalternatives to Chomsky’s narrative. So the Stuff of Thought (which I haven’t read) sounds infinitely more compatible to CL.

    Will follow your review with interest. Keep up the good writing.

  2. Because children can’t learn rules (unless those rules are somehow inborn). This seems a bit like William James’ “turtles, turtles, turtles all the way…”


  1. Nudge Idea: Reinforce foreign language vocabulary with (amost) no effort | Charlie's Blog - To Notice and to Learn
  2. Pinker on Language and Thought (Part II) | Charlie's Blog - To Notice and to Learn
  3. How I Blog a Summary of a Nonfiction Book | To Notice and to Learn
  4. Li-Bary Book: Steven Pinker – The Language Instinct 1/28/16 | As I Please

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s