Seminar Topic number 2

Human Language on the communication continuum:

language features and language functions revisited

1. The properties of language

Given the fact that the ability to communicate is not endemic to humans (which by way of
consequence turns language into a rather fuzzy concept) researchers have always been
interested in pinning down the traits of natural language. Attempts in this direction have mainly
involved the comparison with nonhuman communication modes (e.g. the dance of the bees,
songs, sounds, movements of birds and mammals, etc.), the ultimate goal of which being,
conceivably, to highlight the superiority of human, natural language over all the other forms of
communication.
2. The Design Features Approach
One of the most remarkable attempts to elucidate the problem of what makes human language
special was made by the American linguist Charles Hockett. Two of his articles are particularly
relevant here: The Origin of Speech (1960) and The Problem of Universals in Language. The
Search for Universals Through Comparison with Animal Systems (1966); here Hockett puts
forward the notion of design features, i.e. features/properties present in some communication
systems and absent from others, which, together, constitute a taxonomy by means of which all
communication systems could be described and classified. In its final version, this taxonomy
consists of sixteen design features which separate human language from other or, as the case may
be, from all the other communication modes. Hocketts standpoint is phonocentric, which is duly
reflected in his classification: six of his design features are sound-oriented, i.e. the Vocalauditory channel, Broadcast transmission and directional reception, Rapid fading,
Interchangeability, Complete feedback, and Discreteness.
An overview of these design features together with their glosses would be useful here:

Design feature

Gloss
This feature refers to the vocal-auditory, visual-tactile, or
chemical-olfactory channels in which communication can
occur. As far as human language is concerned, the default
mode of communication is the vocal-auditory channel
since,

in

typical

situations,

messages

are

transmitted/perceived as vocal signs. All known natural

Mode of Communication

languages are basically spoken (i.e. occur in the vocalauditory channel). As such, non-vocal sign systems like
written language or sign-language are seen as more
marginal manifestations of language proper: not all
spoken languages are also written (cf. oral cultures), while
sign languages are only used by (or in communication
with) sensorily-impaired individuals.

A consequence of the nature of the sound itself, this

feature refers to the fact that, when produced, sounds are
Broadcast transmission and
directional reception

sent out in all directions; by way of contrast, their

reception is much more limited: the area within which the
transmitted sounds are audible is more restricted, e.g. the
sounds coming from the front are heard better than those
coming from the back.

Rapid fading (Transitoriness)

Interchangeability

Complete feedback

Once produced, sounds/signs do not linger in time and

space; they fade away quite rapidly.

Individuals within a given community can send and

receive the same signals/messages.

Individuals can hear their own sounds/messages and

correct them in case of error.

This refers to the fact that the human body is anatomically

adapted to the task of speaking the mouth, the lips, etc.
Specialization

are designed primarily for the use of speech rather than

for other functions, like chewing or echolocation. Further,
language signs are specialised since their actual function
is to convey messages.

The signals which make up the communication system

represent/correspond to objects and features from the
Semanticity

extralinguistic reality. In other words, signals have

denotation.

Arbitrariness

Unless iconic or onomatopoeic, signs are completely

unrelated physically or otherwise to the real objects
they represent. As far as natural languages are concerned,
the vast majority of linguistic signs are arbitrary (cf. 2.3.2)

The messages delivered in a language are made up of

smaller, repeatable units. These units preserve their
individuality in the sense that they are discontinuous (i.e.
Discreteness

they do not merge with other units in the same context),

which makes them recognizable in whatever context they
may

occur.

Since

both

symbolic

and

iconic

communication systems consist of discrete units, they are

described as discrete systems.

This feature refers to the potential that human language

has to refer to spatially and/or temporally remote events
Displacement

and situations; it is due to this feature that humans can do

more than just describe the here and now of their
situation; they can recall, anticipate, hypothesize, imagine,

plan, etc.
This basically refers to the ability that language affords
humans to combine and recombine its units in order to
generate novel messages which the other language users
will, despite their novelty, comprehend. It is by virtue of
this ability that language users can be creative
linguistically: not only can they generate novel messages,
Productivity
(Creativity/Open-endedness)

Cultural transmission
(Tradition)

but they can create new words, add novel meanings

to/change the meaning of the already existing ones, etc.
Productivity in a language accounts for its constant
growth and renewal. But innovations in a language are not
random occurrences; on the contrary, they tend to comply
with the pre-established lexical, grammatical, and
phonemic patterns underlying language. This shows that
linguistic creativity is, to all intents and purposes, rulegoverned (cf. Chomsky, 1957).
While it may be perfectly true that language is an inborn
human faculty, language conventions are learned by
interaction with other users of the same language. This
very likely separates the human species from the other
species which do not learn the communicative
behaviour, but are born with it.
This feature can be regarded as a direct cause of linguistic
creativity. It refers to the ability of language to combine
its units on two distinct levels, one meaningful, the other
meaningless. The former consists of meaningful units, i.e.

Duality of Patterning

words/morphemes which combine into larger, message-

(Double Articulation)

carrying structures; the latter includes the meaningless

(but meaning-differentiating) sounds (phonemes) which
combine into words/morphemes. The advantage of duality
is manifest in the economy of linguistic material: it is by
means of a very small set of phonemes (a few tens of

them) that a natural language can create a very large

number of words; these can further be used to encode a
virtually

unlimited

number

of

messages

(q.v.

Productivity above).

Closely connected to Displacement (q.v. supra), this

feature refers to the potential of messages encoded in a
Prevarication

language to be meaningless, deceptive, or untrue.

According to Hockett, the ability to lie appears to be a
typically human feature.

This refers to the metasemiotic function of signs, i.e. their

Reflexiveness

potential to describe themselves. It translates as the ability

of natural language to have its own metalanguage.

Learnability

The user of a natural language is able to learn another

natural language.

If the presence of Hocketts design features is taken as a matter of degrees rather than absolutes,
it becomes obvious that very few of these properties are endemic to natural language. This is
why the validity of this taxonomy has been challenged: it discriminates too little, critics say,
between human and nonhuman modes of communication. Moreover, the approach itself is
objectionable: using human spoken language as a platform of reference puts all the other
communication modes at a disadvantage (cf. Malmkjr 2002: 13).
From among his design features, Hockett himself singled out productivity, displacement, and
duality as the key properties of human language. Besides these, Yule (1996) also shortlists
arbitrariness, discreteness, and cultural transmission which he also sees as core features of
human language. However, insights from research in zoosemiotics and zoopragmatics (cf. Nth
1990:157-164, Malmkjr 2002: 10-16) indicate that the presence of such properties can be
found, to some extent, in animal communication: for example, displacement and productivity
(among other features) can be found in the language of the bees, while some researchers claim to
have identified a form of duality in the song of some bird species.

Generally speaking, judging by the staggering amount of evidence on animal communication

modes, we may come to realize that human language is not as special as we might imagine; or,
that we ought to study animal communication more closely in order to gain a more in-depth
understanding of what makes human communication different; because although we can
acknowledge the differences, it appears difficult to specify precisely what is special or unique
about language (Snowdon 1999:113).
For this reason, Yules (1996) caveat regarding the limitations inherent to the outsiders
perception of animal language is worth keeping in mind:
We should remain aware, however, that our view of how other creatures communicate is
essentially an outsider's view and may be inaccurate. It is possible that your pet has quite
complex communication with other members of its species and frequently comments on
how hard it is to get points across to the large clumsy bipeds who act as if they know it
all. (1996:20)
Much in the same vein, DAndrade (2002) comments on the biased, discriminatory view of
nonhuman communication in the following terms:
On evolutionary grounds, one assumes that different species of animals have just as
much language as they need, and if they somehow had a human speech device grafted
into their brains and larynxes, they would only say what they already communicate quite
effectively. My cat, for example, would say, "Let me out . . . . Well, I don't know if I want
to go out. Maybe I want to come in. No, I want to get out. Well, I will just sit here while I
decide, so leave the door open." In fact, this is what my cat does say without words.
Animals already say quite well all that they have to say; it is just that we have not
specialized in listening. Because they do not use arbitrary symbols as their primary units
of communication and do not specialize in representative speech acts, we self-servingly
deny that they have language and thereby deny that they, too, are saying something [].
(2002:229)

Reverting to the design features of language, it should be noted that Hocketts model has met
with criticism mainly on account of its failure to draw a clear dividing line between human and
nonhuman languages. Could this be due to the fact that there isnt any such hard-and-fast line to
draw? The empirical data available seem to indicate that the arguments in favour of the
uniqueness of human language are tenuous since, as Snowdon points out,
[i]t is easy to find examples of birds or mammals in which all of [Hocketts] features can
be observed. For example, the oft-cited alarm calls of vervet monkeys (Cercopithecum
uelhiops) [] are auditory with broadcast transmission, rapid fading, interchangeability,
complete feedback, specialization, semanticity, arbitrariness, and discreteness. At present
it is an open question whether the production of calls is transmitted through learning, but
vervets' appropriate usage and response to calls does fit the definition of tradition [].
(1999:79-80)
It is also interesting to note that the author uses this empirical evidence on animal
communication to buttress the continuity theory regarding language evolution:
I know of no single nonhuman species that incorporates all of Hockett's criteria in its
communication system, but it should be evident that each of the criteria can be found
either realized or as a potential capacity in some nonhuman species. If one accepts
Hockett's design features as sufficient criteria for language, then clear continuities
exist between humans and nonhuman animals []. [emphasis added] (1999:83)
In the spirit of the continuity theory, built on the belief that the difference between human and
animal language is of a quantitative rather than qualitative nature, Snowdon (1999: 112-113)
hypothesizes on the factors which may have determined the rise of a more complex language
with the human species: the mixture of all the design features, the greater brain power which
allowed for the development of the linguistic ability, or perhaps the propensity/need for social
living, these elements may be taken, the author argues, as points of discrepancy between human
and nonhuman language.
With all the debates it stirred in its wake, the design features approach provides a useful
paradigm within which the (specific and non-specific) features of human language in general and
of natural language in particular can be explored and better understood. Whilst it may have
limited reliability in discriminating between human and nonhuman language, Hocketts design

features allow us to zoom in on the properties of language (in the widest sense of the term) and
thereby gain a clearer view of its essence.
3. The Functional Approach
As an alternative to Hocketts design features, a function-based parallel between human and
animal language has been put forward (cf. Liebermann in Malmkjr 2002). In essence, the
philosophy behind the functional approach is that communication systems are best distinguished
in terms of what they actually enable their users to do, rather than on the basis of their traits.
However, there is at least one underlying assumption which the functional approach seems to
share in common with the design features approach: the superiority of human language. In this
connection, the premise is that
[al]most all animals emit sounds or make gestures in connection with some of these
functions. But humans habitually talk about numerous other subjects arguably,
language has many more, and much more complex, functions than animal
communication systems (in so far as we understand the functions of the latter). The
question is, then, whether this multifunctionality of human language vis--vis animal
communication systems arises because humans are further along the evolutionary path
than animals, and have, therefore, developed far more complex social groupings than
animals, which (in turn) make increasingly complex demands for a communication
system to serve many more and more complex functions (the functionalist explanation);
or whether human language is, in fact, unique to humans and, as such, dependent on a
faculty of the human mind reserved for humans alone. The testing ground for this
question has been experiments intended to teach higher primates to use human
language. [emphasis added] (Malmkjr 2002:13-14)
In day-to-day existence, language may be employed to accomplish various things; we can
therefore speak of a number of pragmatic functions (cf. Austin 1962, Searle 1975) which
various speech acts in a language may perform and which, by virtue of their ubiquity, may be
considered language universals (cf. also 2.4). They have been shortlisted and briefly explained in
the table below, without taking into account their different subtypes:

Function

Directive

Gloss
This function refers to the use of language in order to determine
(or prevent) a course of action. The directive function of language
is manifest when we order, command, entreat, beg, suggest, etc.
and can be expressed directly, as in Sit down, Pass me the salt,
please, What time is it? or indirectly, e.g. Why dont you sit
down?, Would you mind passing me the salt?, Have you got the
time? It therefore involves a message negotiated between an
unstated, implicit subject/sender and a recipient; these two
(sender-receiver) may sometimes coincide in instances like Why
dont I take a break?).

Expressive

This function is manifest in speech acts whereby the language

users intentionally or unintentionally express/vent/communicate
some feeling or state of mind, rather than convey information;
expressives can be either direct, e.g. ouch!, yuck!, damn,
wow, or indirect, as in Its a glorious morning, isnt it? or you
might have asked me first, Im so sorry. Like with the previous
function, expressives also presuppose a recipient, who may or may
not be part of the immediate context of situation.

Whenever speakers use language in order to report on/convey

information about past/present/future events involving someone or
something, they make use of the representative function of

Representative

language. Unlike directives, where both sender and recipient can

remain implicit, representatives are comparatively more specific as
to the who-did-what-to-whom; in other words, given the
temporal and spatial displacement of the events reported, the
circumstances of the event need a more explicit representation in
the sentence. This explains why language instances performing
this function tend to be more complex in terms of their lexicogrammatical structures.

Commissive

This pragmatic function refers to the language used by speakers in

order to make explicit or implicit promises or express various
degree of commitment to a certain course of action. Commissives
can be stated in positive or negative terms, and typically occur in
response to/in conjunction with some other pragmatic function,
such as the directive function. Common instances of commissive
acts are You shall have your money by tomorrow, Ill bring the
wine, Ill do my best, It will never happen again, etc.

Declarative

Seen as illustrations of a basic pragmatic function, declarative

speech acts are associated with some sort of special, (quasi-)
ritualistic/symbolic conditions in which a particular language
instance is accompanied by actual performance of an event. Such
language instances are I declare you man and wife, Youre
fired, I name this ship The Mayflower, I vote no, Motion
carried, We pronounce the defendant guilty, etc.

Directive and expressive uses of language are found in both human and nonhuman
communication; they are, respectively, the first to emerge in the child language acquisition
process. The representative function, on the other hand, has been considered typically human,
due to its reliance on symbols, as DAndrade (2002) points out:
Representation of nonimmediate or displaced events presents a greater demand for true
symbols, for it is difficult to use pointing - a clear indexical sign - to indicate what is not
there. Development of representatives, especially representatives about nonimmediate
events or displaced reference, would have greatly increased the need for the development
of grammar and true symbols []. (DAndrade 2002:225)

Moreover, the need for representative (or representational) language is supposed to have fuelled
the development of human language, which, in turn, has increased our species chances of
survival in the natural selection race:

To account for the development of human language, one must account for the
development of representative speech acts that have the capacity to represent the world
outside the known here and now. For this to have happened, there must have been a point
in human evolution when producing representations of things outside the known here and
now began to be advantageous for individual reproductive success. It is likely that this
would occur with the development of a cultural way of life - that is, with the need to
coordinate learned routines with complex environmental and social information about
things and events not present. (DAndrade 2002:26)
It has been argued, however, that nonhuman communication is also apt to represent displaced
events, i.e. signaling information about the location of prey or feeding places, the approach of
predators, etc., although in such messages the representative function is combined with the
directive function.
Commissive and declarative speech acts appear to be human-specific, given their abstract,
symbolic essence. They rest upon the representative function and conceivably must have
succeeded it in time, since both promise and declaration are based on something one is first able
to represent (DAndrade 2002). It may be inferred, then, that these last two functions might
constitute the differentia between human and nonhuman language.
Considering the framework suggested above, to accept that all the possible things that language
users can do by means of language fall neatly into these five categories may, to some, be a leap
of faith. Moreover, in some cases the clear delimitation of concurrent pragmatic functions is
rather problematic. It will be agreed, however, that this approach yields useful insights into what
is considered to be the specificity of human language and, implicitly, of human language faculty:
grammar and the ability to create and operate with symbols. Indeed, it appears that the
propensity for symbolic signs makes our species stand out from all the other language (in the
broadest of senses) using creatures. It can be argued that using symbolic signs presents us with
the immense benefit of offering shortcuts to communication by allowing us to map what is not
immediately present in either space or time. But if symbols are so handy, why are we alone in
using them?

4 Summary
Unit Four has been premised on the belief that the unique properties of human

language can best be identified if human language is integrated into a wider

paradigm

of

communicative

systems.

This

implies

departing

from

the

anthropocentric stance which holds, a priori, that human language is superior to all
other forms of communication and determining those language properties which
may be taken as endemic to human communication by identifying the features that
different languages hold in common.

Therefore, in this Unit we have been looking at two analytical frameworks

illustrating two perspectives (feature-based vs. function-oriented) from which
commonalities and particularities of human and non-human language can be
considered. What emerges from the foregoing discussion is that both approaches
take human language as the de facto reference system; both suggest that human
natural language is superior to any other form of communication, and the results of
both seem to indicate that when it comes to language, the differentiae between the
human and nonhuman species are rather slim.
Overall, the rather difficult question which this Unit has aimed to answer is What
distinguishes human, natural language from non-human communication?; this, in
its turn, leaves us with another, equally puzzling question: do such distinctions
necessarily attest to the superiority of human language?

5. Exercises
Below you will find a brief list of self-assessment questions (SAQs) intended to guide you
through the information in this Unit; at the same time, they will enable you to test your level of
understanding of the input provided above. Read each SAQ and write down your answer, even if
tentative. Then check your answer against the references to the relevant sections, supplied next
to each SAQ. If this does not help you find the answers or if you are unsure of their accuracy,
bring this up for discussion in the next tutorial.

Exercise 1: Read carefully the following description of the language of the bees. Based on the
information provided, complete the blanks (1 through 5) with the appropriate design features
from Hocketts framework:

A simplified account of the system might go something like this: a bee that
has located a food source will return to the hive and inform its colleagues
of the discovery by dancing to them. If the food source is more than 50
metres away from the hive, the bee dances in a figure of eight, a dance
called the waggle-dance. The length of the straight runs of this dance, up
the long lines of the figure eight, called the waggle-run, is proportionate to
the distance between the hive and the food source, and during the wagglerun the dancer shakes its tail with a vigour which is in proportion to the
richness of the food source. The frequency with which the bee dances also
indicates distance: a bee returning from a food source 100 metres from the
hive dances 10 times every 15 seconds, while a bee returning from 2000
metres away dances only 5 times every 15 seconds. The direction of the
food source is given by the orientation of the waggle-run. If the food
source is less than 50 metres away from the hive, direction is not indicated,
and the bee dances a round dance, which is more lively the richer the food
source. Bee dancing has (1) ______, (2) ______, and (3) ______ of the
type that allows for infinitely many messages to be created, although not of
the type that allows for making new messages of old bees probably only
ever dance about food, not about food as a symbol of anything else. As far
as the workers are concerned, the system also has Interchangeability, and,
in so far as the bee is aware of what it is doing, the system has Complete
Feedback, Specialization, and Semanticity. It does not have (4) ______:
bee dancing is a continuous system because of the proportionality of the
signal to richness and distance of the food source. It is doubtful whether
one would want to claim Tradition for it, and it has no Duality of
Patterning. Nor do bees appear to engage in Prevarication, and there seems
to be no Reflexiveness in the system. Finally, other bees do not learn to
dance like the worker honey bee, so there is no (5) ______ .
(in Malmkjr 2002: 12-13)

Exercise 2: Read the following report (in Yule, 1996) on an incident involving two female
chimpanzees named Matata and Lorel. What design feature is illustrated by the report below?
How does the account below disprove Hocketts initial assumption?
Matata returned to the social group for breeding and found herself
subordinate to Lorel, a female she had easily dominated in earlier years.
The situation lasted for some days, until Matata happened to be alone in the
outer cage with Lorel and the child of a still more dominant female. Matata
reached up and yanked on the child's leg where it dangled on the net above
her. The little chimp squealed, of course. All the other animals came
pounding out of the inside cage, including the adult male and the child's
bristling mother. As they emerged Matata glared at Lorel and barked. The
dominant mother swung round and attacked innocent Lorel. From that day
on, Matata again lorded it over Lorel whenever there was food to take or
babies to groom.
(in Yule, 1996:26)
Exercise 3: Section 3 above ends in an open question. The following text suggests two possible
answers. Read it and decide which answer appears to be more plausible. Think of possible
reasons for your decision.
Most linguists consider true symbols and grammar to be the defining and
essential marks of language. This seems unnecessarily restrictive. That
said, there is something useful about the question "Why don't other
creatures use true symbols?" Deacon (1996) has argued that it is not
convincing to say that symbols are not found in the communicative
systems of others animals because they do not need them, for to whatever
degree they need a communicative system of any kind, symbols could
theoretically be useful.
Deacon's argument is that the reason only humans use arbitrary symbols is
that, given the structure of the primate/ mammalian brain, learning
arbitrary symbols is an extremely difficult task, requiring something
primates just do not do well. The problem the primate brain has in learning
symbols, according to Deacon, involves not just learning that a sign means
that one should do something with some object but learning relationships
among the symbols themselves. The human brain has been uniquely
adapted to do this kind of learning, but if the learning is developmentally
delayed, it can be quite difficult for even a human to learn (Deacon 1996).
(in DAndrade 2002:225)