Speech production

Speech production involves at least two types of process. On the one hand, words have to be selected. On the other, they have to be inte­grated into the syntax.

Slips of the tongue - cases in which the speaker accidentally says something such as par cark instead of ‘car park' - provide useful clues to these processes, and so do pauses: they can tell us where a speaker stops to think - though it is difficult to separate out pauses caused by searching for lexical items, and pauses due to syntactic planning.

There are two main kinds of slips: on the one hand, there are selec­tion errors, cases in which a speaker has picked out the wrong item, as in:

Please hand me the tin-opener (nut-crackers). Your seat's in the third component (compartment).

On the other hand, there are assemblage errors, cases in which a correct choice has been made, but the utterance has been wrongly assembled:

Dinner is being served at wine (Wine is being served at dinner). A poppy of my caper (A copy of my paper).

At first sight, such slips may seem haphazard and confused. On closer inspection, they show certain regularities, so much so that some people have talked about tongue slip 'iaws' - though this is something of an exaggeration. We are dealing with recurring proba­bilities, rather than any real kind of ‘law'.

Selection errors usually involve lexical items, so they can tell us which words are closely associated in the mind. For example, peo­ple tend to say knives for 'forks', oranges for 'lemons', left for 'right', suggesting that words on the same general level of detail are tightly linked, especially if they arc thought of as a pair. Similar sounding words which get confused tend to have similar begin­nings and endings, and a similar rhythm, as in antidote for 'anec­dote', confusion for ‘conclusion'.

These observations were possibly first made by the two Harvard psychologists who devised a now famous 'tip of the tongue' experi­ment, first carried out over 25 years ago. The experimenters

— 129 —

assembled a number of students, and read them out definitions of relatively uncommon words. For example, 'A navigational instru­ment used in measuring angular distances, especially the altitude of sun, moon and stars at sea'. Some of the students were unable to write down the word sextant immediately. The word was on the tip of their tongue, but they could not quite remember it. Those in a 'tip of the tongue state' were asked to fill in a questionnaire about their mental search. They found that they could provide quite a lot of information about the elusive word. They could often say how many syllables it had, what the first letter was, and sometimes, how it ended. They could think up similar-meaning words such as astrolabe, compass, and also similar-sounding words such as secant, sexton, sextet. This suggests that adults store and select words partly on the basis of rhythm, partly by remembering how they begin and end.

A considerable number of selection errors tend to be similar both in sound and meaning, as in component for 'compartment', geraniums for 'hydrangeas'. This suggests that an interactive acti­vation theory, of the type proposed for speech recognition, may also be relevant in speech production. The mind activates all similar words, and those that have two kinds of similarity-, both meaning and sound, get more highly activated than the others, and so are more likely to pop up in error.

Whereas selection errors tell us how individual words are stored and selected, assemblage errors indicate how whole sequences are organized ready for production. For example, mistakes nearly always take place within a single 'tone group' - a short stretch of speech spoken with a single intonation contour. This suggests that the tone group is the unit of planning. And within the tone group, items with similar stress are often transposed, as in

A gas of tank (a tank of gas).

Furthermore, when sounds are switched, initial sounds change place with other initials, and final with final, and so on, as in

Reap of hubbish (heap of rubbish).

Hass or grash (hash or grass).

All this suggests that speech is organized in accordance with a rhythmic principle - that a tone group is divided into smaller units (usually called feet), which are based (in English) on stress. Feet are divided into syllables, which are in turn possibly controlled by a biological 'beat' which regulates the speed of utterance. The

–130–

interaction between these rhythmically based tone groups and syntactic constructions is a topic which still needs to be carefully examined.

Slips of the tongue are part of normal speech. Everybody makes them. But they overlap with the stranger and more extreme errors found in people suffering from speech disorders.