Procflyctoon" putting words 5 mi one's moyth
We are quick to recognize the exceptional precocity of talented writers, artists, or athletes, but we often fail to appreciate the gifts underlying so many of our everyday activities. It is only through loss or injury that we suddenly realize how much we take them for granted. The skill involved in such a literally pedestrian activity as walking down a flight of stairs is immediately recognized after one has sprained an ankle. It is only then that we begin to appreciate the marvelous manner in which the visual input from our eyes and the tactile information from our feet transmit complementary information to our brain's sensory cortex. There it is immediately synthesized and fed to corresponding areas of the motor cortex which, in turn, feeds the cerebellum, the part of the brain devoted to the programming, timing, and coordination of all voluntary muscular movements. From the cerebellum radiate hundreds of simultaneous messages along the nerve pathways which go to the appropriate muscles involved in the head and neck (to focus the face and eyes downward toward the stairwell), in the back (to keep the posture erect and tilted slightly backward to compensate for the downward motion of the body), in the arms and hands (to slide down the banister for continual support and feedback), in the legs (to maintain a lifting and dropping motion quite different from normal walking), and in the feet (to angle the foot in just the right manner so that the ball of the foot catches the stair). Even this elaborate description is a gross oversimplification of the neurosensory and neuromuscular processes that are involved at any single moment of a descent down a staircase. But all of this is taken for granted and considered uninteresting, until we stumble and injure ourselves. Loss of what we consider the
-i (\ qiipvcv
simple and common gives us renewed appreciation of life's uncommon complexity.
The production of speech is neurologically and psychologically far more complicated than negotiating a flight of stairs, but its intricacy also goes unappreciated until we suffer some linguistic disability or commit a slip of the tongue. In daily conversations, we remain generally unaware of the complexity of our achievement. Again, it is only through disability that our marvelous ability is made manifest.
We have already seen in the previous chapter that psycholinguists tend to divide linguistic phenomena into stages. One of the most influential psycholinguists models for speech production, developed by Levelt, views it as a linear progression of four successive stages: (i) conceptualization, (2) formulation, (3) articulation, and (4) self-monitoring. We will look at each of these in turn, not forgetting that viewing speech phenomena as a step-by-step sequential process is only one way of investigating production. Alternative approaches exist; for example, characterizing the production of speech as a holistic activity where several simultaneous and parallel activities are taking place to create the utterances we intend to produce.
Conceptualization
Where does the very beginning of any spoken utterance come from? What sparks speech? These are difficult questions to answer, partly because we still don't know enough about how language is produced, but partly because they deal with mental abstractions so vague that they elude empirical investigation. The American psycholinguist David McNeill, however, has gone on record with an interesting mentalistic account of how speech is first conceptualized in the human mind. His theory is that primitive linguistic concepts are formed as two concurrent and parallel modes of thought. These are syntactic thinking, which spawns the sequence of words which we typically think of when we talk about how language is initiated, and imagistic thinking, which creates a more holistic and visual mode of communication. The former is segmented and linear and creates the strings of syllables, words, phrases, and sentences that together make up speech.
The latter is global and synthetic and tends to develop the gestures which we naturally use to punctuate and illustrate our conversations.
McNeill's claim, that syntactic thought and imagistic thought collaborate to conceptualize conversation, is quite convincingly demonstrated by the way in which speech utterances and ordi- 1 nary gestures seem to be tied and timed together in any conversation. Consider the following very simple example. Two people are holding a short discussion over the whereabouts of a lost object. Visualize in your mind how they gesture as they interact in the following two dialogues. You might even try reading these aloud, acting out Person B's role by pointing at the appropriate moment.
First dialogue
Person A: Where's my briefcase? Person B: There's your briefcase!
Person B points to the briefcase the same moment he says There's.
Second dialogue
Person A: Where's my coat and briefcase? Person B: There's your briefcasel
Person B points to the briefcase the same moment he says briefcase.
What are the very first things that are going through Person B's mind when she is responding to Person A's questions in these two dialogues? Of course we cannot be too mentalistic and pretend we know what B is thinking. After all, we are often unsure of what we are thinking ourselves when we think about what we think, if we think about thinking at all. This is the problem with mentalism. But McNeill offers some plausible evidence for this bimodal view of how speech is produced. It seems likely that after B hears A's query in the first example, her syntacric thought might generate something that begins with the demonstrative, 'there' while, simultaneously, her imagistic thought might be of someone pointing toward an object, in this case, a briefcase. Evidence that these two modes are operating concurrently at the conceptualization stage is found in the simultaneous timing of the pointing gestures with the stressed words in each of these two scenes. In the
first dialogue, B points to the briefcase (manifesting the imagistic part of her attempt to communicate) just as she stresses the word 'there' in her speech (illustrating the syntactic component of her communicative intent). Again, in the second dialogue, we see the synchrony of image and speech; at the end of the phrase B points to the briefcase just as she stresses the word in her articulation. If you read this last example out loud, you will also note a slight change in B's intonation—the voice trails off a bit as if to say 'There's your briefcase ...' Were B suddenly to spot the coat, she could continue with 'and there's your coat\ with a more decisive, falling intonation on 'coat' and, of course, another pointing gesture to show A where his coat was located.
Appealing as McNeill's hypothesis might appear, and convincing as these examples might be, it is difficult to use his model to explain this first stage of production. For one thing, his attempts to describe how imagistic and syntactic thought are initially conceptualized are unclear. For another, the illustrations he uses to describe how gestures synchronize with important syntactic breaks in spoken language are difficult to follow. Perhaps this form of research, like studies of American Sign Language, can only be adequately illustrated by a videotape and not by drawings.
Levelt's initial stage of conceptualization seems justified. After all, speech does not start from nothing, and if it does not start with concepts, how else could it possibly begin? At the same time, we realize how difficult it is to actually define this stage in non-mental-istic terms, and despite the plausibility of McNeill's binary model of language and gestures being birthed together, like twins, it is difficult to muster any hard evidence to support this, or any other theory for the embryonic development of speech. Although we know very little about how speech is initiated at this first stage of conceptualization, we have psycholinguistic evidence to help us understand the successive stages of production, so it is easier for us to describe and to understand Levelt's second stage, formulation.
Formulation
Introduction
We have seen that the initial stage of conceptualization is so far removed from the words we actually speak and write that it is
difficult to delineate this phase of production. But at the second stage of speech production, formulation, we move close enough to the eventual output of the process to allow us to be more precise in our terminology and more convincing in our use of empirical data. Conceptualization is hard to conceptualize, but formulation is much easier to formulate. Well over three decades ago, the psychologist Karl Lashley published one of the first attempts to account for the way speakers sequence strings of sounds, words, and phrases together so rapidly and accurately, and his essay was influential enough to be included in the first book ever published in English which focused exclusively on the then very new field of the psychology of language. His essay was first presented as an oral address, and it is intriguing to see how Lashley organized it to demonstrate some of the very concepts about speech production which he was writing about. For example, he talked about how common it is to commit spelling errors when one is typing, and he mentioned how he misspelled 'wrapid' with a w, while typing 'rapid writing', most probably because as he was about to type 'rapid', he anticipated the 'silent w' in the following word. These slips of the tongue, or pen, or computer keyboard, are of keen interest to us in this chapter on production. A moment later in his talk, to illustrate several of the themes that were central to his presentation, Lashley gave the following utterance as an example of how we comprehend spoken sentences.
Rapid righting with his uninjured hand saved from loss the contents of the capsized canoe.
Remember that this sentence was beard, not seen, so having been primed by the earlier phrase 'rapid writing', it was natural for the audience to hear 'Rapid writing with his uninjured hand!' Of course, like all native speakers of any language, the listeners were able to readjust their comprehension of this sentence. After they recognized they had initially wandered down the wrong garden path of comprehension, they were forced to retrace their steps, and to choose the proper path toward complete understanding. Thus Lashley was able to demonstrate many of the themes which were central to this seminal essay on speech production. First, he showed how slips of the tongue (or the computer
keyboard) provide vivid insights into our understanding of how speech is formulated. Second, he illustrated the power of priming in guiding the direction of speech production and comprehension. Because Lashley first talked about 'writing', his audience was primed to hear the phrase again, and this is what confused them initially when they heard 'rapid righting' as part of an utterance about a canoe. Note also that it is possible that priming works in the production process as well.
A critical insight from this example, and from Lashley's essay as a whole, is the way it demonstrates how both the production and the comprehension of speech is largely a linear process. The audience didn't know that Lashley had purposely misled their comprehension until they suddenly heard something about 'saving the contents of the capsized canoe'. People tend to produce and comprehend sentences in a linear way, and for comprehension, each additional piece of information we receive has the potential to force us to revamp our understanding of what we have already heard. The comprehension of the 'canoe' example, taken from Lashley's lecture and subsequently published paper, lends credence to the notion that, in several ways, production and comprehension are similar. Both are largely sequential, both are affected by priming, and they both depend, to a large degree, on the constant winnowing of implausible alternatives at each juncture in our stream of speech.
Slips of the tongue
Think back to the example at the beginning of this chapter of how we tend to ignore the complexity of strolling down a flight of stairs until we trip. Over the past few decades, psycholinguists have become excited about a new way of discovering how we put words into our mouths: they look at what happens when we trip over our tongues. Unlike stammering or aphasia (linguistic loss due to brain damage), slips of the tongue, or typographical mistakes, are normal, everyday occurrences which pervade our speaking and our writing. And because of this, as soon as our friends spot our mistake, or we happen to catch the goof ourselves, we can immediately backtrack and correct. However, when speech and language disintegrate into clear pathologies, as they do in stammering and aphasia (to be discussed in Chapter 5),
there appears to be no recourse, and the error remains uncorrectable and uncorrected. So we can see why slips of the tongue provide the data that delight psycholinguists; they allow us to peek in on the production process because we know what the speaker intended to say, but the unintentional mistake freezes the production process momentarily and catches the linguistic mechanism in one instance of production.
The use of linguistic deviations as data for scientific investigation is a new phenomenon, but the recognition of speech errors goes back more than a century. Spoonerisms, like the unfortunate use of 'the breast in bed' instead of 'the best in bread', are named after the Victorian cleric and teacher, William Spooner, who reputedly blundered through many a lecture or sermon with infamous slips in speech production. He called a group of Welsh miners 'you noble tons of soil' and supposedly scolded an errant student by saying, 'you have hissed all my mystery lectures; in fact, you have tasted the whole worm!' Spoonerisms then are slips of the tongue in which an actual word or phrase is created, often with a humorous twist to the meaning which was intended.
The mention of verbal miscues, especially ones like the 'bread' example just cited, evokes a discussion of Freudian slips. In one of his earliest treatises, Psycbopathology of Everyday Life, Sigmund Freud hypothesized that slips of the tongue were important because, like dreams, they help to reveal the unconscious mind. But most psycholinguists have ignored Freudian interpretations of speech errors for a variety of reasons. For one thing, although slips like 'the breast in bed' appear to be embarrassingly indicative of an unconscious desire, a coldly empirical approach to this mistake would propose at least two explanations for the source of this illicit feeling: either the speaker was sexually provoked (and was unconsciously thinking about the first noun in the phrase), or he was actually fixating on the second noun because he was so exhausted! Here, we run into exactly the same problem we faced earlier when we tried to define what the conceptualization stage for speech consisted of; we are in danger of becoming too mentalistic—that is, relying on logic and intuition rather than experimental evidence—in our attempts to fathom what exactly puts words into people's mouths. A more important reason why psycholinguists tend to ignore Freudian analyses of why who said
what is because, irrespective of their meaning, the formulation of slips of the tongue reveals important linguistic patterns—patterns that also pervade normal and natural speech. That is, what is relevant to psycholinguistics is not what is being said, but how it is being said, or, to be precise, misspoken.
But before leaving Freud, it might be illuminating to examine some recent work in the psychology of language that attempts to wed the Freudian, mentalistic tradition of psychology with the experimental school which so strongly colors contemporary psycholinguistics. This was the goal of an experiment in which university students were asked to read aloud two, unrelated words flashed quickly in front of them on a computer screen. For example, the subjects might have seen 'barn door' instantaneously flashed in front of them, and they either read them correctly, or, as was often the case, because of the pressure of time, came up with a slip of the tongue, such as 'darn bore'. There were three different situations: a control—or normal—situation, where the subjects had no distractions; a second situation where subjects knew that they might receive a small electrical shock at any moment, and a third situation, where they performed the task in a state of slight sexual arousal (the subjects were all male, and the experimenter was an attractive and well-dressed female). Here are examples of two stimuli phrases which were flashed to all three groups.
(i)sham dock      (2) past fashion
Although most of the subjects were accurate most of the time under all three conditions, the slips of the tongue which did occur differed significantly among the three groups. The control group tended to make arbitrary errors, such as 'darn bore', but the two experimental conditions tended to elicit two different kinds of slips. When (1) was flashed to the subjects who were in the group that was threatened with a potential electrical discharge, they, much more frequently than the other two groups, came up with the slip 'damn shock'. And when (2) was shown to the group with the attractive female experimenter, they, as you have already anticipated, came up with the phrase 'fast passion' much more frequently than the others. This experiment comes about as close as we can expect to get to testing Freud's ideas under laboratory
SURVEY
conditions, or to catching a glimpse of the conceptualization stage of speech production. Be that as it may, our focus here is on formulation, and from all of the examples cited so far, we can readily see that slips of the tongue are not a random, haphazard zigzagging of the production mechanism, like quarks in a cloud chamber. Sounds and words are not thrown together arbitrarily; there is a clear, linear and hierarchical order to the way in which we put words into our mouths.
There has been a long and rich tradition of examining speech errors in psycholinguistics as a window to the formulation process and not as a reflection of some Freudian motivation. Based on examples gleaned over the years, researchers have been able to demonstrate that these superficially trivial quirks of communication are quite useful in offering insights about how speech is formulated. For one thing, there is sure evidence from this data that the units of speech, such as 'phoneme' and 'morpheme', which linguists have proposed and discussed for many years are psychologically real. This means that when we misspeak, we make errors within the boundaries and the framework of a certain language structure, as if we had intentionally planned our slips to fit an appropriate linguistic slot. Mistakes do not pop out just anywhere when we speak, they occur at predictable points and follow predictable patterns. It is almost as if we think about syllables, words, and phrases as we are formulating what we are going to say, and this is why psycholinguists find slips of the tongue insightful. Let us review some of this evidence.
Linguists divide sounds into vowels and consonants and sub-categorize each of these into various phonetic groupings. Speech errors seem to follow the phonetic classifications established by linguists and rarely, if ever, cross over these linguistic boundaries. Consider the following examples.
(1) a reading list   a /eading list
(2) big and fat     pig and fat
(3) fill the pool     fool the p/'ll
(4) drop a bomb   bop a dromb
As trivial and silly as these mistakes may appear initially, they actually tell us a great deal about the organization of the English
language. The anticipation of [1J in the third word in (1), creates the substitution of [1] for [r] in the second word. Phoneticians point out that [1] and [r] are two consonants which share many-phonetic features for example, both are pronounced in the same part of the mouth, so that this type of substitution would always be likely. There is no such phonetic explanation for a substitution of [1] for [sh] {e.g. 'shopping list' becoming 7opping list'), and, in fact, whereas flip-flops of [1] and [r] for each other pervade the miscue data, transpositions of [1J or [r] for sounds like [sh] are exceedingly rare. The second example is a bit more subtle, because at first sight it seems that [p] is randomly introduced into the phrase from nowhere. But again, linguistic analysis gives a clear explanation. The phonetic feature of voicelessness of the following [f] in fat seems to be anticipated when the speaker is about to produce the [b] in the first word of phrase (i). As it turns out, the voiceless equivalent of the consonant [b] is [p]. Although we are focusing on sound structure in these first few examples, it is also possible that speakers are simultaneously being influenced by other linguistic factors, so that the person who misspoke 'pig and fat' may have also been gently nudged by the semantic association between these words.
In (3) we see the psychological reality of the contrast between vowels and consonants in the minds of speakers. The vowels in the two words replace each other. It is theoretically possible for vowels to substitute for consonants and vice versa, but again, this rarely occurs because they are so distinct linguistically. Finally, in (4) we see how speech errors follow rules about what consonants can go together to form clusters. We can't just put any two consonants together in English, or in any other language for that matter: although we have [dr] in 'drop' and [st] in 'stick', we have no words that begin with [sr] like 'srop' or even worse, [dr| as in 'dtick'. So even though there is no word 'drom' in English, the [dr] cluster that the slip in (4) creates is permissible, and it tells linguists that people who come up with odd expressions like these still follow the sound patterns of English.
Slips of the tongue also reveal that when we formulate speech, we are not only influenced by the sound system of the language we are speaking, we are also conditioned by the way words are put together in that language. Consider the following examples as
evidence of the psychological reality of morphology—the way words are organized and structured in a language.
(5) sesame seed crackers sesame street crackers
(6) rules of word formation words of rule formation
(7) a New Yorker a New Yorkan
(8) the derivation of the derival of
Unlike the first set of examples, these slips do not involve individual sounds; rather, they seem to reflect a higher level of linguistic organization because they are associated with complete words, or with significant parts of words. (5) and (6) are very common examples, and they remind us of the spoonerisms discussed earlier. Notice how the misspoken forms still adhere to normal patterns of word usage. For example 'Sesame Street crackers' might be a brand of cracker named after the children's TV show. Note too the manner in which (6) adheres to a regular word pattern in English. A 'four-door sedan' has four doors, an 'apple pie' is made of apples. A common error by learners of English is to call these objects a 'four-doors sedan' and 'apples pie', following the logical, but non-English pattern of extending the plural to the formation of noun phrases. But native speakers, who follow the rules of word formation, do not simply swap the two words that are reversed in (6) and say 'word of rules formation'. Even during the micromomentary process of formulating their speech, they follow the regular and established pattern.
Examples (7) and (8) are further elaborations of this same theme, but in this case, the suffix slots are exchanged while the original words remain the same. The person who misspoke (7) might have been thinking, if an 'American' is someone who lives in America, why isn't a resident of New York a 'New Yorkan?' And by the same logic, if 'arrival' is the noun form of the verb 'arrive', why isn't the noun form of 'to derive', 'derival?' Once again we witness the way slips of the tongue provide psycholinguists insights into the production of speech; they help us see how speakers arrive at derivations.
Speech errors are also helpful in revealing a third level of language processing at the formulation stage; they give support to the notion that utterances are not just strings of sounds and linear sequences of words, but are formed into larger structural units.
This is demonstrated in examples (9) and (10).
(9) he swam in the pool he swimmed in the pool
(10) the children are in the park   the cbilds are in the park
These mistakes are much less common than the swapping of words and parts of words that we find in spoonerisms and similar constructions, but their occurrence, however rare, tells us something about the way grammar affects the formulation process. Those familiar with the speech and writing of non-native users of English will recognize these goofs as learner errors, but the big difference between learner errors and the slips exemplified by (9) and (10) is that native speakers almost always correct themselves when they err; learners of English, on the other hand, experience great difficulty recognizing exactly what was wrong and how to rectify it. Almost immediately after saying (10}, for example, a native speaker might stop and say 'I mean children''. Learners, upon recognizing that they said something wrong in a sentence like (10), or, more commonly, having it pointed out to them, will often miscor-rect the original error and come out with something like 'I mean cbildrens\ The fact that native speakers correct themselves shows that they are also paying attention to grammar, in addition to concentrating at the sound and word levels of the language. It is no accident that these last examples all involve irregular words. It looks very much as if the speaker has chosen the words and the slots which they fill, and at the last moment, forgotten to choose the right verb or noun form. The errors suggest that speakers organize their utterances into smaller groups of words, like noun phrases, or clauses with a main verb, and having filled these groups with the appropriate lexical items which express the intended meaning, the speakers finally add the appropriate grammatical inflections. Almost always, this complicated process is completed fluently and accurately, and only occasionally, as in these examples, does the formulation of speech slip up. But when it does, it provides us with a glimpse of the production process.
The planning of higher levels of speech Another way of trying to understand the process of producing language is to analyze the steps we have to take and the decisions we have to make in order to produce an intended utterance.
3<i STJRVFY
r
Suppose, for example, that you might want to give a response, either spoken or written, in a certain situation. Let's say you are in a discussion with a friend ahout the importance of a particular matter, and your friend asks for your opinion about the gravity of the situation. You decide to frame a response, and for whatever reason, you choose to conceptualize the idea that the matter was not important. How do you go ahead and formulate this concept linguistically? Granted, you are constrained by all the phonological, lexical, and grammatical patterns of English which were exemplified and supported by the slip of the tongue data just discussed, but, despite all these rules and patterns, there is still a great deal of flexibility in what you say and how you say it.
Of course, how we choose to formulate what we are about to say or write is influenced by such factors as politeness or social appropriateness. These extremely important variables are not usually dealt with in the relatively asocial circles of psycho-linguistics, but they are central to the concerns of linguists who investigate pragmatics—the study of what people mean when they use language in normal social interaction, or those who study sociolinguistics—the study of why we say what to whom, when, and where. But just within the narrow confines of how we formulate a simple concept such as 'not important' into actual words, following the rules of the language we have chosen to speak, and disregarding all the complicated nuances of pragmatics and sociolinguistics, we still have many choices to consider. Given all the decisions we are forced to make every time we begin to open our mouth, it is quite astonishing that conversations aren't painfully slow trickles of syllables, dripping intermittently from tongue-tied interlocutors. In fact, they are often the reverse—cascading torrents of speech pouring so rapidly and so easily that conver-sants overlap and interrupt each other in their eagerness to fill the silence.
We will freeze this fictional conversation at the very instant when one speaker decides to put the concept 'not important' into words. Many choices come to mind. Should the concept be expressed lexically, that is through the choice of a word, or should it be expressed by means of a syntactic pattern? Supposing the speaker makes the first choice, further alternatives still remain. Will the word chosen be grammatically negative, as in
(a), or affirmative? If the latter, will the word be an antonym, or opposite of 'important' as in (b), or will it contain an explicit negative prefix like un- in (c)?
(a) It's nothing.
(b) It's trivial.
(c) It's uttimpbrtant,
It may seem a bit peculiar to differentiate among these three choices by calling the first negative and the other two affirmative, because they all convey the same idea, that the speaker believes the situation is not important. But language is not always as it seems, a fact that most native speakers are blithely unaware of and one that allows psycholinguists to make a living. Think, for a moment, about one very common way of asking questions in English—by placing a tag at the end of an utterance. You've probably asked questions like this a lot, haven't you? Observe that an affirmative sentence, like the tag question immediately preceding this one, uses a negative tag, and vice versa, so if we choose to formulate a negative sentence, then the tag is affirmative. You're not confused about this, are you? Let us go back to our three examples and see what happens when we transform these sentences into tag questions.
(d) It's nothing, is it?
(e) It's trivial, isn't it?
(f) It's unimportant, isn't it?
The affirmative tag 'is it' in (d) tells us that even though there is no overt negation in this short response, it is still considered grammatically negative. If it weren't, it would have a negative tag at the end, wouldn't it? In contrast, both (e) and (f) must have negative tags because they are both grammatically affirmative phrases. Nevertheless, they are not similar: (e) expresses the concept of 'not important' by choosing an antonym; (f) says the same thing by using a negative prefix with 'important'. And the choice of prefix in ff) is an additional complication, because English proffers a wide range of potential prefixes. Some words take several choices (for example (/^European, «o»-European, anti-European), but even words like 'important', which only take un-, create a production problem. When speakers choose 'trivial',
the only thing they have to remember is the word itself, but when they select a word which takes a negative prefix, they have to recall which one to use. Again, we often fail to realize the complexity of the production process until we see it fall apart, as in the ///correct choice of a prefix by a learner (for example (///possible and cf/simportant).
Let us go back now to the initial choice the speaker made. Recall that the first alternative the speaker had in formulating the concept 'not important', was whether to express the negative response lexically—via words, or grammatically—via the use of syntactic negation. Let us suppose the second alternative was picked, creating the series of choices exemplified by sentences (g) and(h).
(g) It isn't important.
(h) It's not important.
You may have to glance at these twice to catch the slight differences between them, and they are so minimal that you might be provoked into using some of the earlier examples: the differences are nothing; they're trivial! But if the speaker has chosen to express the negation grammatically rather than through word choice, important differences can be indicated by means of stress. Normally, the contracted negative in (g) is chosen because negation is typically not the focus of our attention, but (h) offers an effective way of emphasizing negation. Supposing you are in an argumentative state, and your conversational partner keeps insisting that the situation is desperate; (h) allows you to be emphatic about your denial. Put tersely, the difference between the two sentences is that in (g) the negative is not usually stressed, but in (h) it receives unusual stress.
The significance of these slight differences may seem minimal within the context of the myriad sounds, words, and sentences that comprise our daily staple of communication, but along with the slip of the tongue examples, they demonstrate the enormous number and intricacy of choices facing a speaker, or a writer, at this important stage of formulation.
Articulation
We have spent considerable time examining the second stage of speech production, and for good reason. Like the operation of a computer program during word processing, the formulation stage of speech involves thousands of split-second decisions regarding the hierarchical and sequential selection of myriads of potential segments. But this third stage of articulation is similar to what happens when all of those bits of information selected by a word processing program go from your computer to your printer; unless this vast amount of electrical data is 'articulated' into letters of the alphabet and successfully printed, no message is received. In fact, if the printer is not functioning properly, there is no evidence that the message was ever even composed. So, too, with the production of speech. Unless all of the electrical impulses streaming from your brain in the form of speech are transformed into audible and comprehensible articulations, no words are heard and nothing is communicated. The conceptualization stage might pompously perceive itself as the primary and ultimate composer of communication, and the formulation stage might pride itself as the conductor and orchestrator of speech sounds, but without the instruments of articulation, the music of our voices remains unheard and unappreciated. Like the operations of a printer or the playing of instruments then, the articulation of speech sounds is a vital third stage of production and, quite naturally, attracts the interest of psycholinguists.
As recently as the 1960s, linguists upheld the common sensical and seemingly incontrovertible notion that the chest, throat, and mouth were anatomical organs designed solely for biological functions. Only in a secondary way could they be considered the organs of speech. Surely, the basic function of our lungs is to exchange oxygen for carbon dioxide, not to produce syllables and, most assuredly, the primary use of our teeth is for chewing, not for the articulation of sounds like [t] and [d]? True as these assertions may be, they do not preclude the possibility that some organs may have been shaped in their recent evolutionary history to enhance the production of human speech sounds. Some thirty years ago Eric Lenneberg, a psycholinguist, showed that whereas the majority of these organs have primarily evolved to serve
essential biological functions such as respiration and ingestion, a few of them have adopted secondary functions connected with the enhancement of speech articulation. In a few cases, there are organs that have changed anatomically to fit this new role as speech articulator. So dramatic are the changes that they differ physically from the corresponding organs in closely related species like chimpanzees.
Perhaps the most dramatic example of an organ which has adapted itself for human articulation is the larynx—the 'voice box' which houses our vocal cords. Like all the other speech organs, the larynx did not initially evolve with the specific function of helping humans to articulate language. For one thing, the vocal cords in all animals possessing a larynx serve as a kind of emergency trap door which can prevent foreign matter, such as bits of food, from falling from the mouth down the pharyngeal tube and through the trachea into the lungs. When bits of debris do manage to find their way down these passageways, the vocal cords help control explosions of air from the lungs to cough this potentially life-threatening jetsam back up out of the mouth. The larynx thus helps keep the respiratory tract clear, but it serves another primary function which is just the opposite of coughing. By squeezing tightly closed, it can trap air in the chest cavity and create a solid fulcrum for the limbs to work against when heavy physical exertion is required. However, in the case of our species we could claim that speech has now become the primary function of the larynx and the other, original purposes of the voice box have diminished to secondary stature.
Evidence for the evolutionary modification of the human larynx to create speech is quite dramatic. Lenneberg and others have documented several speech-enhancing characteristics of the voice box that are unique to humans and are absent in other mammals, even the primates like chimps and gorillas. No wonder then that they have remained unable to master articulate speech. The most striking difference between humans and all other animals in this area of the body is the position of the larynx. In all other animals, the larynx is found high in the throat, crammed behind the tongue, an exceedingly advantageous position for preventing debris from entering the trachea, for it can be trapped immediately as it leaves rhe mouth. But this is not true for us. Feel your
neck and find your larynx (or 'Adam's apple1). You will locate about halfway down, almost touching the top of a high-collared shirt or blouse. Consider the awful consequences of this anatomical deviation from natural evolution. Unlike all other animals, our emergency trap door cannot stop foreign matter as soon as it leaves the mouth. It is so far down that a passageway, called the pharynx, has been created into which debris can easily fall, and we, among all creatures, are the most susceptible to choking. Why does nature deviate in this destructive manner for humans?
The advantages of the lower voice box reside in the way this j. arrangement serves to embellish the articulation of speech sounds. Unlike other mammals whose highly-positioned larynx virtually precludes the existence of a pharyngeal tube linking the back of the mouth with the opening of the vocal cords, the pharynx benefits the production of speech in at least two ways. It creates a new source of speech sounds—the 'throaty' consonants of Arabic, or the initial consonant of the two words in the English salutation, 'Hi Harry!' A pharyngeal tube also increases resonance by adding extra acoustic space to the already existing oral and nasal cavities. The addition of a pharynx to the vocal repertoire is not unlike the addition of a cello to the duet of a violin and viola; the timbre of the human voice is that much richer, thanks to the added instrument. Another enormous benefit of the lowered larynx is the way it frees up the back of the tongue so that the tongue root can maneuver and create more speech sounds. The contrast between the vowels in the words 'look' and 'Luke' is made largely by subtle movements of the tongue root, movements that no other animals are capable of performing. So the linguistic advantages outweigh the physiological disadvantages, and if the emergence of language is as vital to our evolutionary history as mosr anthropologists believe, and if language is so indispensable to our species, it is no exaggeration to claim that the descent of the larynx has permitted the ascent of mankind!
Given the anatomy of articulation we have been endowed with, what do we know about the programming of articulation? How do sounds trip so miraculously off the tips of our tongues once speech is conceptualized and formulated? It is easy to assume that speech sounds are produced in a linear, sequential fashion, like cars off an assembly line, but a closer analogy might be to the
r
team effort that goes into producing a batch of cookies. While one person might be chopping the walnuts, another might be preparing the cookie dough, while a third might be preheating the oven and greasing the cookie sheets. So too in the production of speech sounds. The process might appear to be linear, but the lungs, larynx, and lips may be working all at the same time, and coarticula-tion is the norm, not the exception. That is, in the production of any single sound, a lot of anatomical effort is devoted to performing several different movements simultaneously.
Consider just one sound, the second [k] (spelled, of course, with a 'qu') in the expression: 'Keep qiriet kid!' Let us begin by contrasting the second [k] with the first (in 'keep') and the last (in 'kid'). All three of these sounds are dorsovelar which means they are made in the back of the mouth. The back of the tongue (the dorsum) hunches up and touches the soft palate at the back of the mouth (the velum) to stop the flow of air momentarily to produce the consonant [kj. But the process is much more complicated than this, because every sound in the stream of speech is affected by the sounds which swim around it. None of the [k]s in this particular phrase is preceded by any sound which would demonstrably affect its pronunciation but each is followed by sounds which do affect articulation. The first and last are followed by a vowel pronounced in the front of the mouth, so the [kj in 'keep' and the [k] in 'kid' slide forward a bit from their usual position in the back. It is almost as if the tongue were at the starting line of a sprint and was trying to inch up a bit to get a head start in the race to those front vowels. There is a double contrast between the [kj in 'quiet' and the first and the last [k]'s. First, because the vowel in this word at least begins with a sound in the back of the mouth (the initial [a] of the diphthong [ai]), the [k] in this word does not inch forward at all, but actually sits well back in order to hit the following back vowel. Secondly, because 'qu' is actually a cluster [kw] and not a single sound, the initial [k] is rounded in anticipation of the following [wj. In other words, the lips assume an 'o' position when we begin to articulate 'quiet', whereas they remained flat and unrounded in the production of the first and last words. Try reproducing this phrase in front of the mirror, and you'll obtain visual evidence of the effects of coarticulation; the mouth momentarily puckers up when you begin to pronounce
ctinifrv
'quiet'. Here we observe the complexity of articulation. Sounds do not emerge as segments strung together sequentially; they are mixed and melded, with each sound shaping its neighbors while concurrently being shaped themselves.
Psycholinguists have developed a number of competing models to try to account for the complexity of speech articulation, and they have tried to employ various sources of evidence to peek into this complicated process, but much of articulation remains a mystery. For example, despite the increasing sophistication of modern neurology and the development of techniques such as Positron Emission Tomography (PET) scans to examine the way the human brain programs neuromuscular movements, we still have little understanding of how the cerebral software programs the anatomical printer to articulate sounds in such a glib manner. Let us narrow the issue down to one simple question. How does the tongue 'know' to cheat a little bit ahead in the first and last words of the phrase described in the paragraph above, but 'know' not to inch forward in the second word. And how do the lips 'know' when to pucker up? It would be impossibly difficult to explain the rapidity and accuracy of articulation in such closely related phrases simply as a chain of habits acquired in a linear way.
We see then that even at this seemingly uncreative and mechanical aspect of speech production, complexity and mystery abound, and speaking ceases to appear a simple and mundane act. Speech production does not end with articulation, however; the fourth and final stage of production is t]ie process of self-monitoring.
Self-monitoring
Earlier, during our review of slips of the tongue, it was noted that the production process sometimes goes awry and speakers will verbally misstep, especially with irregular or more unusual forms. Almost always, however, they instantly catch themselves, retreat a step, and correctly recreate the intended sequence, as in (i) and U).
(i) The last I knowed about it {I mean knew about it], he had left Vancouver.
(2.) She was so drank (I mean drunk], that we decided to drive her home.
In contrast to the conceptualization and formulation stages of production but similar to the articulation stage which we just reviewed, it seems that at this final stage of self-monitoring we have direct evidence of what is happening when people compose speech. Interlocutors not only produce speech and listen to one another when conversing, they also seem to keep one ear open on what they themselves are saying, and if they catch something amiss, they are quick to amend the goof and then continue to converse.
All speakers and writers of any language, regardless of their degree of native fluency, commit linguistic blunders. To err is human. Common to all speakers too is the way in which fluency seems inversely proportional to the amount of attention they pay to the production process. It also appears to vary inversely in proportion to the degree of stress they are under, or the quantity of certain beverages they have imbibed! S. Pit Corder, a pioneer in the field of Second Language Acquisition (SLA) classified these slips of the tongue and the pen as mistakes. The examples listed earlier in this chapter, or goofs like the typos and misspellings we are all responsible for from time to time, are mistakes, whether they come from native or non-native mouths and fingers. Mistakes are production problems; they are the troubles you have with your linguistic printer, not with the original sofrware. The true test of a mistake is to see whether or not it is corrected, and (i) and {2) above, as well as most of the other illustrations we have covered, are surely mistakes under this definition. Errors, on the other hand, are committed only by non-native speakers (NNSs) according to Corder. Even when the speech gaffe is pointed out to NNSs, they have difficulty correcting it. For example when NNSs goof up in the production of irregular verbs, as beginning learners of English often do, they frequently fail to notice that they erred in what they said or wrote. Even when they are told that they erred, unlike the native speakers exemplified in (1) and (1), NNSs do not immediately replace the deviancy with the correct form.
The fact that native speakers do not commit 'errors1 (in Corder's sense of the word) coupled with the fact that they often produce
'mistakes' which they almost immediately self-correct, reveals three insights into the production process. First and most transparently, it demonstrates that speakers (and writers) are constantly self-editing. Production is not a one-way transmission of messages; it is a self-regulating process with a feedback loop to ensure that each previous stage of output was accurate. Second, it suggests that speakers are intuitively sensitive to what stage of the production process went awry, if indeed a mistake was made. Speakers and writers are quickly capable of readjusting a message at the stages of conceptualization, formulation, or articulation, depending on where they noticed the breakdown in production occurred. Finally, the fact that native speakers can monitor and quickly correct any mistakes in linguistic output proves Chomsky's contention that there is a distinction between performance and competence. The former refers to the words we say or write, the overt manifestation of our ability in a language; the latter describes our tacit, intuitive knowledge about the language or languages we have mastered. At this final level of production, competence monitors performance to ensure that our production is accurate.
There are several different ways native speakers edit their linguistic performance. Psycholinguists have studied the kinds of self-repairs speakers make and have discovered some significant contrasts in the ways they monitor their output. When speakers were dissatisfied with the social or situational appropriateness of their speech (i.e. their choices were wrong at the very beginning, at the conceptualization stage), they were much more likely to backtrack and begin the utterance all over again; however, if speakers were content with the conceptualization of their utterance, but had somehow goofed up at the formulation or articulation stages, they were less likely to start afresh. They would retreat a few syllables or words to the linguistic juncture in the utterance where plans began to fall apart, and renew the sentence from that point, as seen back in examples (1) and [%},
Once more we witness the way in which language and speech provide windows to human cognitive processing. Evidence appears strong that stage models of production, such as those posited by Levelt and described in this chapter, accurately reflect the manner in which people produce and edit conversations or compositions.
Along with mistakes, such as slips of the tongue, psycholinguists have also relied on the trivial and sometimes annoying hesitations which punctuate our unplanned, spoken discourse to gain insights into the ways in which we monitor the language we produce.
(3) I think it costs just about... lib ... twenty-five dollars.
(4) They have to try to ... uh ... contact an attorney.
Hesitations like those exemplified in (3) and (4), or the ubiquitous 'y'know' which pervades a great deal of contemporary conversation, are not mistakes—certainly not in the sense that the term has been defined and illustrated here. Nevertheless, they do seem to indicate a lack of fluency. But however clumsy they might appear to an articulate speaker of the language, they are not random, but rule-governed, and hence they are of interest to psycholinguists. Notice how the hesitations in (3) and (4) appear at a crucial point in the sentence—before the object of the verb in (3) and before the complement of the verb in (4). The intrusive 'uh' suggest that the conceptualization phase is still in the process of selecting the information to appear at the end of the sentence, after the verb in these examples, and so the speaker pauses in mid-utterance to allow the computer to program the last part of the message to be printed. However awkward these hesitations might sound, they reflect the grammatical structure of the language being spoken, and they would almost never violate linguistic constraints. We very rarely find utterances like (5) and (6).
(5) I think it costs just... uh ... about twenty-five dollars.
(6) They have ... uh ... to try to contact an attorney.
'Just about' and 'have to' function as linguistic units, so it is improbable that the speaker would hesitate in the middle of either one, after having already chosen to fill the linguistic slot of the utterance with those phrases.
One final point about self-monitoring. It contrasts markedly with the dated and very inaccurate depiction of communication as consisting of a message that speaker A sends to listener B. The attested presence of a self-monitoring stage presumes that people don't just communicate with others, they communicate with themselves; they don't just listen to others, they listen to them-
selves. Communication is not a one-way broadcast of a signal, but it is an interactive process, involving not just the interaction between the interlocutors but also the interaction within each individual speaker. The self-editing process confirms for psy-cholinguistics what has long been known to exist in most biological functions of the body—the presence of feedback loops. Speech production (or written composition} is not a linear 'one-way' process; it is a parallel, 'two-way' system involving both output and the concurrent editing and modulation of that output.
Let us go back to the theme with which we began this chapter. All of the complexities of production which we have reviewed here are largely overlooked by speakers. Language generally flows effortlessly, and even our hesitations, slips, and backtrackings are so swiftly executed that they go mostly unnoticed. It is only when this marvelously evolved and efficient instrument of communication breaks down that we appreciate its intricacy. And it is also only then that we begin to glean significant psycho-linguistic insights.