Five Basic Language questions (v4)

1. What is language (L)?  (Description of structures)

An utterance in a language can be described at various levels.  For example when I say 'stop', a linguist might describe the relevant body movements, the acoustic product (waveform), and the linguistic descriptions that apply – phoneme, syllable, morpheme or word, phrase, clause, sentence, semantic representation.   The overall meaning of the utterance is built up out of the meaning of the individual morphemes in their syntactic and pragmatic context.

It is important to convey the idea that language is more than a bunch of words – that may be true for apes and aphasics.However fluent language users have layers (levels) of abstract hierarchical "tree" structure mediating between movement and meaning.

(See Language video notes.)

2. What are the functions of L?

The obvious answer is that L serves as a means of communicating information between individuals.  Most uniquely, this information is not restricted in regards to content or specificity of reference.  (Maybe see Pinker (1994) or Hockett's 1959 paper on 'design features' for ideas.)

Linguists use expressions like 'using finite means to express infinite ideas.'

It is useful to look at functions from three different perspectives – interpersonal communication, intra-personal functions, and social functions.  Important examples of intra-personal functions might include memory aids – internal and external, as well as reflective consciousness.  Social functions include defining group memberships via accents and transmitting precise cultural information enabling an accumulation of knowledge unlike that seen in any other species.

It is very unlikely our language skills enable us to introspect directly into most of our cognitive processes – keep in mind the interpreter/confabulator that creates stories for any occasion.

To sum up, human language serves as more than a primate interpersonal communication adaptation. It is likely several of these functions drove its evolution. Darwin's (1871) comment on brain-language co-evolution is worth repeating (though ignore the Lamarckian overtones!)

"As the voice was used more and more, the vocal organs would have been strengthened and perfected through the principle of the inherited effects of use...but the relation between the continued use of language and development of the brain, has no doubt been far more important....we may confidently believe that the continued use and advancement of this power would have reacted on the mind itself, by enabling it and encouraging it to carry on long trains of thought." (Origins, 1871)

3. How does L work?  (Processes)

Much remains unknown here; moreover there are several perspectives that one can take – speech production and speech perception – that involve the same structures but quite distinct processes.  (Advocates of a motor theory of speech perception have tried to reduce these into more or less one problem; this may gain new life with the discovery of 'mirror neurons'.  See recent PBS video.)


Our auditory system transforms movement generated time pressure waves into neural impulses.  These waveforms are 'parsed'  -- given a linguistic description resulting in a semantic representation for the utterance in terms of agents, actions, and objects.  At the very least, we automatically understand who did what to whom at the level of sentence structure.  Information from our lexicon, current context and grammatical knowledge provide us with the knowledge linguists represent in a complex 'tree' diagram.

This is only the beginning of a complex process.  Some idea of the complexity can be glimpsed using examples that can be parsed in more than one way, i.e. given two structural descriptions. (See old handout and my 1976 paper online.)

Pragmatics and reference

The final interpretation of the utterance is an interaction of this semantic representation with the context – the current situation and what Pinker calls the listener's 'vast database' of knowledge.

(See my examples including Bransford & Johnson, estimation of height .)


In fluent speakers, our thoughts are encoded into movements at a rapid rate – we can easily speak at 200 words per minute and up to twice that if necessary.

Hundreds of decisions are made each second – what morphemes, what syntax, how do those choices interact in the linear speech output. etc?  The 'yolk'-priming priming demo suggests the complexity and some of the variables in speech production. Much of the time, I don't really think I know my own thoughts until I get the corollary feedback – the English-speaking voice in my head.  This gets into lots of issues about the relationship between our thoughts and our language.)

4. How is L acquired? (Acquisition)

Many of us who have studied human language acquisition have concluded it is a biological process enabling children to rapidly acquire the language artifacts of their culture.  Each of the levels (above) must be acquired within the first 2 to 3 years, with continued development of fluency for a decade and growth of vocabulary throughout life.

Typical acquisition

Human language is acquired by all humans, without tutoring or explicit rewards and it occurs rapidly and with great predictability across a wide range of environments.

What do infants expect?

Linguists have worked to determine what aspects of language are universal – that is are common to all human languages.  Some of these universals are presumed to be part of the 'Language Instinct' and need not be learned from scratch.  Phonological structure, hierarchical phrase and sentence structure, as well as lexical structures are shared across all human languages.

Creole formation, in both spoken and sign languages, demonstrates how children exposed to only fragments of language will create their own.

The mere availability of language to children, however, does not imply they use their language as adults do.  (Consider various examples including Simcock & Haynes (2002), and others we have discussed.)

Almost every word we use is a result of exposure to that word; its origin however is often unknown but we can be sure it was invented by someone else – the process of lexicalization.

  Sentences are quite the opposite – we invent them as we need them out of existing words using general syntactic rules.

5. How did L evolve?    (Evolution)

The only plausible answer here follows the one Darwin suggested – language evolved from the cries of animals.  Yet human language is so distinctive in its discrete combinatorial processes and hierarchical phrase structure that this takes us nowhere.

At best, we can take a look at several issues and hope for the best!

Animal communication


All mammals use their larynx for a variety of communicative functions – maternal-infant, mating, aggression, group control, territory claims etc.  Humans seem to have acquired a great deal of  voluntary control over the breathing and vocal apparatus compared with many species, especially chimps.

Non-vocal communication may involve other body movements, olfactory and coloration cues.


Birds of course have complex calls; many are learned.  Others have great mimicking ability like parrots.  Many have lateralized control of vocalization suggesting one reason for lateralization – lack of conflict in rapid variation of movements.

The biology of bird song learning is fairly well known and interesting  involving sex-differences, new neurons holding newly learned songs, and innate templates filled with personal experience.

Human-based codes with animals

This deals with the training of animals to use 'human' language fragments.  In every case, this is an exaggeration.  Be wary of claims here; the signs taught to Koko are not American Sign Language;  Kanzi is not at a seven year old's language level.  (see my old paper on this; little needs to be revised but I would elaborate on Alex and Kanzi.). An organism knows a human language to the extent it knows and can use, the structures of that language.

On the other hand, there is real communication that allows a certain insight into the animal mind.  Pepperberg gives a good account of her parrots and the whole issue.

Evolution of homo Sapiens

The one origin theory

This is essentially the idea that humans evolved once, less than 200,000 years ago in Africa.  This group emigrated to populate the world, eliminating previous homo species, including Neanderthals.

This ended a process of homo X evolution beginning several million years ago.  The story begins with divergence of apes and homo ancestors 5-7 million years ago.


Shows up in fossils (Lucy) about 3.4 mya.  Recent conjectures suggest the earliest bipedalism may be as ancient as 6 mya. The upright head possibly enabled subsequent vocal tract development, as well as brain size growth.

Rapid brain growth

(PFC, Broca, hippocampus, cerebellum.. relatively larger)

Lateralization of functions

Language and the interpreter – a few hundred thousand years old or less?

Modern vocal tract

The oldest fossil with an apparent modern vocal tract is less than 150,000 years old.  Compare with a chimp. they have far less degrees of freedom of movement, as well as slower rate of movement. See the old Kellogg efforts to train Vicki to vocalize operantly

The human tract gains in vocal control and range, it loses in preventing choking.

Specialized hearing?

In many respects, mammal ears function similarly, with different hearing ranges related to frequency ranges of their vocal tract.  A recent Science paper reported that the right ear advantage (REA) for human speech actually starts at an early point in the auditory system rather than somewhere up in auditory cortex.  I donŐt know if chinchillas also have a REA for speech!

'Language genes'

Every month a new paper discusses chimp-human gene differences, in the context of vocal tract, sensory, or brain genes.

Co-evolution of language and brain?

Did language become an 'instinct' because only those who could catch on fast survived?  (See Limber, 1982 where I discuss the  'Baldwin effect'.)

History of languages

Language families

At present there are some 6000 languages.  This somewhat arbitrary designation can be organized into language families that share a common ancestor.  English belongs to the Indo-European family that includes Slavic, Germanic, Celtic, Latin, and Greek, along with some languages spoken in Persia and northern India including Sanskrit.  Other families include African, Semitic, various Asian and new world languages.

Families are reconstructed primarily on the basis of shared morphemes and phonetic transformations of those morphemes, e.g. compare brother, brat (Russian), frater (Latin).  There is a dictionary based on reconstruction of the proto-Indo-European language spoken some 8000 years ago, more or less in eastern Turkey or western Persia???


The one-origin theory

Just as humans evolved once (if so!), it seems likely that human language evolved once.  From that initial dialect – the mother language! – all other dialects of human language emerged by cultural evolution. 

Early on, this may have shaped the brain to rapidly acquire and effectively use that knowledge (the idea of co-evolution; also the "Baldwin-effect."). For more on language evolution, see my 1982 paper "What can chimps tell us about the origin of Language?

Language is the premier "mind tool". Did it or some primitive version precede or follow other tools?

(from Science, 1988).



Languages change because children acquire their language a bit differently from what their parents had.  The changes may be small from one generation to the next, but over a few thousand years they may be significant – just look at English!

Specialty topics

There are many topics that can be illuminated from the point of view of human language; and conversely.  Some of these might include poetry, linguistic relativity, sign language use, Creole formation, dialect issuesÉ language disorders  (hemispherectomy, aphasias, SLI).

Discussion of specific topics of interest

Each of these is intertwined; perhaps the most important is the first (description of structures; all the others depend on a good answer here.  Earlier conceptions of language as a bunch of words or responses led to bad answers to the remaining questions.

Thanks to modern linguistics, we have a fairly good working knowledge of human language structures.  This is commonly discussed in the traditional terms of linguistic levels of description' or simply 'linguistic levels.'

It's important to point out that these levels of description apply both to sentences and words (morphemes, lexical items).  As Pinker points out, both sentences and words have phrase structure, along with meaning and phonological representations.

We need to say something useful about meaning; I will have to work on this beyond priming demos.

Good demos or other materials

Priming – yolk, Stroop, Bransford & Johnson (laundry)

McGurk effect

Berko's wug test

Demo of speech sound components, pitch and its origin and function


Language, in the Mind Series.  See my notes on this.

The Human Language (3 50 minute parts)