NAME: RUSHABH MADHUKAR GADE

ROLL NO. 071

CLASS. BBA IB

SUBJECT. BCS

TOPIC. PARA LANGUAGE

Paralanguage
Tim Wharton

‘We speak with our vocal organs, but we converse with our whole body’.

(Abercrombie, 1968: 55)

Introduction
According to Albert Mehrabian’s ‘7%-38%-55% Rule’ (1981), 55 per cent of the emotional

information communicated in a given exchange is communicated via ‘body language’, 38 per cent

via the way say the words we say (intonation, affective tone of voice etc.) and only seven per cent

from the words themselves. While the accuracy of these claims has been criticized (Trimboli and

Walker 1987), and, indeed, Mehrabian (2014) himself has gone on record as saying his results are

often misrepresented, the so-called ‘non-verbal dominance view of communication’ continues to

play an important role in ethology, developmental psychology, Freudian psychoanalysis and other

fields. The very existence of such a rule is at least suggestive. When it comes to the communication

of affect, there is an assumption that the non-verbal dimension of human communication is at

least as important as the verbal one, if not more so.

All of which, rightly or wrongly, has been of virtually no concern to linguists. The non-verbal

contributions to linguistic communication were left behind by the so-called ‘cognitive revolution’ of

the 60s and 70s (Hecht and Ambady 1999) as generative linguists abstracted away from

paralinguistic phenomena in order to focus on the linguistic code. Some functionalist linguists,

discourse analysts and sociologists continued to address aspects of non-verbal communicative

behaviour (see, for example, Bolinger 1983; Goodwin 1981; Brown and Yule 1981; Garfinkel 1967;

Goffmann 1964; Gumperz 1971; Hymes 1972; Schiffrin 1994), but distinctions that are important

from a pragmatic view have been left relatively unexplored. The question of how that 93 per cent

of behaviours might interact with the seven per cent is largely ignored.

Those of us who work in pragmatics can afford ourselves no such luxury. Firstly, the aim of a

pragmatic theory is to explain how utterances are understood, and utterances, of course, have

both linguistic and non-linguistic properties. Secondly, the emotional dimension to speaker

meaning is at least as important, often more important, than those dimensions that tend to receive

more attention. Any pragmatic theory worth its salt simply must have a view on nonverbal

communicative behaviours and how they contribute to speakers’ meanings.

A problem of definition
But when it comes to the analysis of ‘paralanguage’, there is a problem of definition. According to

Poyatas (1993) paralanguage includes only the vocal aspects of language use that are not, strictly

speaking, part of language: affective tone of voice, the non-linguistic elements of prosody

(effectively, the 38% in the above rule). Facial expression and gesture, which some researchers do

analyse as paralinguistic (see below), are not part of paralanguage. These, he claims, are better

analysed as elements in a broader category of ‘kinesics’, which includes posture and ‘proxemics’.

Proxemics concerns itself with the way communicators orient themselves to one another and the
distance they maintain between themselves. (We only become aware of this when someone

invades our ‘body space’). Paralanguage is also sometimes defined so as to include ‘oculesics’ – the

frequency of glances, patterns of eye fixation, pupil dilation – and ‘haptics’ – the communicative

domain of touch.

For Poyatas, paralanguage is:

[those] nonverbal voice qualities, voice modifiers and independent utterances produced or

conditioned in the areas covered by the supraglottal cavities (from the lips and the nares to

the pharynx), the laryngeal cavity and the infraglottal cavities (lungs and esophagus), down

to the abdominal muscles, as well as the intervening momentary silences, which we use

consciously or unconsciously…

(1993: 6)

A radically different stance is taken by Abercrombie (1968). He defines paralanguage so as to

include all those aspects of linguistic communication that are not part of language per se, but are

nonetheless somehow involved with the message or meaning a communicator conveys.

In order to avoid too much overlap with other sections of this handbook, the conceptualisation of

paralanguage adopted here will necessarily be the one proposed by Poyatas. But notice that, so

defined, the set of paralinguistic behaviours is a very (very) small one – basically, all those aspects

of prosody that are not deemed grammatical – and so before continuing I would like to add the

following two caveats.

Firstly, and perhaps most importantly, defining paralanguage to include vocal but not, say, facial

expressions is to overlook the vitally important observation that they work closely in tandem with

each other. Consider the following from Bolinger:

This kind of working in parallel is easiest to demonstrate with exclamations. An ah! of

surprise, with a high fall in pitch, is paralleled by a high fall on the part of the eyebrows … A

similar coupling of pitch and head movement can be seen in the normal production of a

conciliatory and acquiescent utterance such as ‘I will’ with the accent at the lowest pitch –

we call this a bow when it involves the head, but the intonation bows at the same time.

(1983: 98)

The parallels are so strong that a single, homogeneous account of these para-/non-linguistic

behaviours seems to be required, one that embraces the fact that they are, for the most part,

closely interlinked.

Secondly, in recent years there has been a huge amount of work in pragmatics suggesting that the

gap between linguistic meaning and speaker meaning is much wider than the one originally

sketched by Grice in his distinction between what is said and what is implicated (1967, 1989). Work

since his famous William James lectures suggests that speaker meaning is not just

underdetermined by linguistic meaning but massively so (Bach 2001; Blakemore 2002; Carston

2002; Sperber and Wilson 1986/1995; Recanati 2004). And that gap is bridged not by more coding,

but by inference.
The central aims of pragmatics are to explain how, by and large, hearers are accurate in their

reasoning (in the sense that it broadly reflects the speaker’s communicative intentions), the

principles behind that reasoning, and the route it takes. Any attempt to characterise linguistic

communication, then, must reflect the fact that it is an intelligent, inferential activity involving the

expression and recognition of intentions. An inferential, not a coding-decoding approach is

required, and our analysis of paralanguage should be no different.

Towards a pragmatics of paralanguage

Paralanguage and pragmatics: the state of play

Many approaches to the broad category of nonverbal communicative behaviours, however, have

been highly resistant to developments within pragmatics. Adam Kendon’s work has provided

numerous insights into gestural communication, as well a wide range of descriptive categories into

which gestures fall (McNeill 1992, Kendon 2004). But he explicitly endorses a view under which

the, as he terms them, invisible, ‘mysterious forces’ of intentionality are beyond the realm of

scientific study (2004: 15). Cooperrider 2011, who largely endorses this view, outlines some of the

arguments against adopting a more ‘pragmatic’ view on gestural communication. In his work on

facial expression, Alan Fridlund (1994) also circumvents entirely the notion of intentionality.

Research into paralanguage is much the same, and works on an assumption dismissed long ago by

those working in pragmatics: that if something is communicated, it must be because it encoded

(Gussenhoven 2002; Ladd 1996; Ohala 1984). Ohala’s ‘Frequency Code’ is a prime example. It is

commonly known that the frequencies produced by small animals, with consequently small

larynxes, are typically higher pitched than those produced by larger animals with large larynxes:

according to Ohala, high pitched vocalisations encode harmlessness, unassertiveness or submission,

and low pitched vocalisations encode danger, assertiveness and dominance. Inference, intentions,
pragmatics are simply nowhere in the picture. Lacking these key concepts, an explanatory account

of precisely how paralanguage contributes to human linguistic communication will remain

incomplete.

MeaningNN and showing

Paul Grice’s ‘Meaning’ (1957) is one of the most influential philosophical papers of the past fifty

years, and remains a huge influence on linguists, pragmatists and cognitive scientists as well as

philosophers. In it, he began by drawing a distinction between what he called natural, or indicator,

meaning and non-natural meaning (meaning NN). The latter he attempted to characterise in terms of

the expression and recognition of intentions. Consider the distinction between the kind of meaning

inherent in (1) and (2):

(1) That black smoke meansN the tyre factory is on fire.

(2) That white smoke meansNN the Vatican Conclave has elected a new Pope. 1

For Grice ‘what is meantNN’ is roughly coextensive with what is intentionally communicated, and his

notion of non-natural meaning has had a major influence on the development of pragmatics. But a

point often missed is that in the course of a communicative exchange paralinguistic phenomena

(some of which are, in a sense, natural) often play a role in determining what has been meant NN.

Consider the following examples:

(3) S (voice trembling): I feel fine, honestly…

(4) S (smiling to A, in a delighted tone of voice): It’s great!

(5) S (lying, deliberately faking a happy tone of voice): It’s wonderful news…
 (6) S: Aaaaaargh! That flaming hurts! Ow! Ow! Ow! Ow! Ow! Ouch!

(7) S (shivering, exaggerating a bilabial trill): Can we go please go inside?

In (3), vocal manifestations of the (S)peaker’s illness indicate to his (A)ddressee that he is nowhere

near as well as he would like them to believe; in (4), S’s delighted tone of voice calibrates just how

great S believes whatever she is referring to is; in (5), S fakes a natural behaviour, which indicates

that he is being ironic, and means the opposite of what he has said; in (6), S’s natural paralinguistic

expressions of pain say as much as the words he utters; in (7), S exaggerates a shiver, perhaps so as

to provide evidence for her request.

A controversial feature of Grice’s account of intentional communication is the line he draws

between showing and meaningNN, where meaningNN typically involves a linguistic convention or

code. This distinction has had important effects on the evolution of pragmatics since, following

Grice, pragmatists have focused on the notion of meaning NN and tended to abstract away from

cases of showing. However, I have argued (see Wharton 2009), that while there is room for

disagreement on whether cases of ‘showing’ always amount to cases of meaning NN, there is little

doubt that cases of ‘showing’ do qualify as cases of intentional communication of the kind a

pragmatic theory should be able to handle: meaning and communicating do not always line up.

Bearing this in mind, and looking once more at examples (3)-(7), but this time through the lens of

an inferential-, as opposed to code-, model of communication, a range of further questions comes

to mind. In some cases, for example, paralinguistic behaviours betray our thoughts and feelings to

others in a way that does not amount to intentionally communicating them (see example (3)). In

others, they may be deliberately produced in a way that clearly amounts to intentional
communication (see example (4)). In still further cases, they are involuntarily produced but may be

deliberately (or overtly) shown (see (5)) and even exaggerated (see (6) and (7)). Intentional verbal

communication, then, involves a mixture of natural and non-natural meaning, and an adequate

pragmatic theory should take account of both. Many aspects of paralanguage are natural but also

shown.

Signs and signals

A further complication is that it appears not all phenomena that mean naturally work in the same

way. In the study of the information transmission between non-human animals a distinction is

made between signs and signals. Signs carry information by providing evidence for it. Signals, on

the other hand, are those natural behaviours that convey information and have been ‘moulded by

natural selection to do so’ (Seeley 1989: 547). It is their adaptive function to carry information,

where the adaptive function of a given trait or behaviour is the effect which is historically

responsible for the reproduction and propagation of that trait or behaviour within a species

(Millikan 1984, Sperber 2007).

Whilst a sign may happen to carry information for an observer, it would go on being produced

whether or not it carries this information. As a result of regular travel across dusty soil, a predatory

species might leave traces of their presence. Certain species of prey might learn to associate such

traces with the predator’s presence. The traces themselves, however, cannot be said to have a

signalling function. Signals, by contrast, do have a communicative function. The function of the

honeybee’s dance is to inform other honeybees about the location of nectar. The function of the

bullfrog’s call is to alert female frogs to the fact that he is searching for a mate. If they did not carry

this information, it would be hard to see why these behaviours survive. Most animal

communication seems to be based on signalling systems of this type. 2

In Wharton (2009) I argued that these distinctions apply in the whole range of human

communicative behaviours. I illustrate the distinction between natural signs and signals in the

human case by comparing shivering with smiling (compare (4) and (7) above). Shivering is a natural

behaviour whose function is to generate heat by rapid muscle movement. It may provide evidence

to an observer that the individual is feeling cold. However, it is not its function to carry this

information: it is not a signal but a sign.