SPEAKING AS A MULTIPLE-LOOP PHENOMENON
The speech system may be viewed as a closed multiple-loop system containing feedforward and feedback internal and external loops. Following are details concerning the internal and external loop aspects.
INTERNAL LOOP
This aspect is concerned with all those processes which may take place within the individual and which are responsible for speech formation and monitoring and speech production and monitoring. Seven processes may be recognized in the operation of this internal system. It should be stressed, however, that not all seven processes are necessarily engaged during any particular speech act.
THOUGHT PROPAGATION. Thought patterns are evoked by either external stimulation or evoked from within, and may take the form of various types of images (although more often than not they take the form of inner speech).
------------------
From Edward Mysak, "Speech System," Speech Pathology and Feedback Theory (Spring field: Charles C Thomas Publishing Co., 1966), 17-36. Reproduced by permission of the author and publisher.
---------------------
2. WORD FORMATION. The feeding forward of developed thought patterns to areas in the brain responsible for creating corresponding word patterns constitutes the word formation process.
3. THOUGHT PATTERN-WORD PATTERN COMPARISON. The thought pat tern-word pattern comparison process describes so-called inner speech where one checks on what one is about to utter. Not all speaking situations require such processing. However, when a speaker desires to utter something in just a certain way, or if one needs to be cautious in his choice of words, this operation may be engaged. Such inner speech monitoring represents one of the feedback loops which make up the internal aspect of the multiple-loop speech system.
4. WORD PRODUCTION. Actual word production occurs when impulses from the word formation area are fed forward, activating the primary motor speech areas which, in turn, innervate the appropriate respiratory, phonatory, and articulatory musculature.
5. ACTUAL WORD PRODUCT-DESIRED WORD PRODUCT COMPARISON. The operation of actual word product-desired word product comparison involves error scanning and measuring of the speech product; this activity ensures the articulatory accuracy of the speech output. Internal loops in this circuit may carry auditory, tactile, and proprioceptive signals back to the brain for processing. Speech product feedback monitoring describes the function of this operation.
6. WORD PRODUCT-THOUGHT PATTERN COMPARISON. Concomitant with speech product feedback monitoring is speech content feedback monitoring. This feedback process ensures as high a degree of correspondence between thoughts and words as may be possible. Toward this end, the individual continually scrutinizes speech content output, compares it with his thoughts, and makes appropriate adjustments when necessary. The operation represents still another aspect of the internal, multiple-loop system.
7. SPEECH RECYCLING. If internal, multiple-loop activity is error-free, there is a continuing reduction of thoughts into words, or speech recycling.
Inspection of this internal, multiple-loop complex reveals at least two levels of activity: A higher "thinking level" devoted to thought and speech content monitoring, and a lower more automatic "doing level" devoted to articulatory (also rate, loudness, and voice factors) or speech product monitoring. It is be cause of these two levels of operation that an individual who misarticulates may not be auditorially aware of his error sound; that is, such an individual may be consciously engaged in speech content monitoring with his ears, but because speech product monitoring is usually left to tactile and proprioceptive channels, and is therefore on a less conscious level, he is not aware of his error sound.
EXTERNAL LOOP
The feedforward aspect of the external loop of the multiple-loop speech sys tem involves the directing of the spoken message at a listener; the feedback aspect consists of the evaluation of listener reactions by the speaker and the making of appropriate output corrections depending on the nature of these reactions. In order to complete the series of operations performed by the internal and external multiple-loop speech system, three more processes must be added to those already discussed. These are: word product-listener reaction comparison; actual listener reaction-desired listener reaction comparison, and, finally, if all is error-free, speech recycling. Further information pertaining to these latter processes will be presented in the next section.
In summary, then, considering both the internal and external loop aspects of the total speech system, the following ten operations may be recognized during a full cycle of speech behavior: (1) thought propagation; (2) word formation (feedforward) ; (3) thought pattern-word pattern comparison (feedback) ; (4) word production (feedforward) ; (5) actual word product-desired word product comparison (feedback) ; (6) word product-thought pattern comparison (feedback) ; (7) internal, multiple-loop speech recycling; (8) word product-listener reaction comparison (feedback) ; (9) actual listener reaction-desired listener reaction comparison (feedback) ; and (10) internal and external, multiple loop speech recycling.
INTERNAL, AND EXTERNAL, MULTIPLE-LOOP SPEECH BEHAVIOR
Next, an example of the operation of a multiple-loop speech cycle will be presented. Let us imagine speaking to an individual and attempting to recall a mutual friend's name: First, an idea is generated which stands for the friend, for example, you may visualize the friend's face or some other characteristic (thought propagation). Second, the idea of the friend automatically (and reciprocally) excites the appropriate word association which, let us say, is "Joan" (word formation). Third, there may be feedback and comparison of the word with the idea or "inner speech" checking (thought pattern-word pattern comparison). Fourth, once it is willed, the release of the word pattern automatically results in the excitation of the appropriate neuromuscular configuration, and the word "Joan" is uttered (word production). Fifth, there is automatic speech product feedback (auditory, tactile, and proprioceptive signals) monitoring or inspection for the articulatory accuracy of the spoken word "Joan" (actual word product-desired word product comparison). Sixth, a concomitant speech content feedback monitoring occurs on an auditory basis, whereby the accomplished spoken word is checked with the idea that it is supposed to represent (word product-thought pattern comparison). Seventh, if the system is free of speech product and speech content errors, it proceeds to process additional speech cycles (internal loop speech recycling). In order to illustrate the error-measuring, self-adjusting nature of the internal loop mechanism, or corrective internal loop recycling, let us suppose that the spoken word "Joan," during the sixth operation, feeds back error signals into the speech system to the effect that the word "Joan" is not the correct name after all.
This causes the system to automatically scan for a new word association for the thought pattern, and let us imagine that these recycling procedures result in the development of the new word product, "Jane." This new word product may now be observed to create speech system stability.
Eighth, the external loop becomes active as this new word product is directed at the listener for his consideration. Suppose the corrected word product "Jane" causes the listener to shake his head in a negative fashion, and that these negative signals are fed back to the speaker. The listener reaction may be due to the fact that the word "Jane" elicits a thought pattern of a person other than the mutual friend in question and this situation causes the automatic generation of negative signals from the listener (word product-listener reaction comparison). Ninth, because the speaker seeks agreement from his listener, he carefully scans the reaction of his listener. And, as an example of external loop error measuring and self-adjusting, the error signals emanating from the listener are received and acted upon causing corrective recycling, and the system may be noted to return to the original word product "Joan." This word product may now pro duce positive feedback signals from the listener (actual listener reaction-desired listener reaction comparison ). Tenth, when there are positive feedbacks from both speaker and listener, the speaker proceeds to reduce new thoughts into words and the conversation continues (multiple-loop speech recycling). The foregoing example reveals that dynamic oral communication depends on the interaction of a series of automatic and reciprocal relationships within the many internal and external loops which comprise the multiple-loop speech sys tem. It also brings to mind the principle of the unity of the multiple-loop oral linguistic circuitry; a principle analogous to that described by Meader and Muyskens (1959) in their discussion of the unity of the organism. Four aspects of the principle may be recognized.
1. RELATIVITY. All parts of the internal and external loop oral linguistic circuitry are interrelated.
2. SELF-REFLEXIVENESS. Every part of the circuitry tends to influence every other part.
3. NONELEMENTALISM. No part of the total circuitry can be fully under stood in isolation.
4. UNITARY STRUCTURE. Understanding of each part of the total circuitry in its relation to the other is essential to the understanding of the total circuitry.
The principle has important theoretical and practical significance to the speech scientist and clinician. So often research or remedial procedures are con ducted without consideration of one or another of its aspects.
Finally, returning to the speaking example, malfunctioning in any one of the ten operations described may reflect itself in some type of oral communicative disorder. Problems may develop from speech generation or feed forward sources, or from speech monitoring or feedback sources. Certain disorders, therefore, are suggested by speech system cybernation which are not now covered by our present systems of classification.
FUNCTIONAL COMPONENTS OF THE INTERNAL SPEECH LOOP
Detailed descriptions of the various functional components included in the internal loop of the speech system as well as the anatomical and physiological representations of these components will now be presented. A model of the internal loop has already been presented by the author in a journal article (Mysak, 1959) ; the model was an extension of one designed by Fairbanks (1954). Com parison of that model and its accompanying discussion with the present Figure 1 would reveal numerous changes. These changes have come about due to the avail ability of new data and also in the interest of simplifying and clarifying the ex pressed concepts.
FIGURE 1. Cybernetic analogue of the speech system.
RECEPTOR
The receptor unit represents the first section of the internal loop. It is made up of three basic components which subserve the estimating function of the sensorium. This section processes sensations such as: radiant energy via the eye (receptor 1), sound pressure energy via the ear (receptor 2), and mechanical energy via the end organs of touch (receptor 3). Proprioceptive end organs (receptor 4) are not included; however, it is possible that proprioceptive sensations may also be utilized in speech reception in special cases. The eye and ear are commonly recognized as speech receivers, but it should be recognized that touch and secondary movement sensations can be associated with meaning also.
For example, individuals without sight or hearing can learn to make meaningful associations by reacting to touch sensations arising from the act of writing in the palms of their hands; or by feeling the movements of the articulators of the speaker as they move and make certain contacts. It has been pointed out that such sensing of articulatory movements, as well as associated intraoral breath pressures, may also have remedial benefits when used with certain types of mis-articulators, speech retardates, and aphasics. If the whole receptor unit were utilized in sending a word to a normal individual, he would receive every sensory dimension of which a word is composed. To illustrate: the word might be uttered while the listener listens to and watches the speaker, while he touches and feels the speaker's associated articulatory activity, and while he has his own articulators moved simultaneously through the various articulatory positions by the speaker.
It may be noted that Figure 2 uses a finger, an eye, and an ear to represent tactile, visual, and auditory reception. The cybernetic analogue (Figure 1) places all the components within one unit to indicate the usual concomitant bi-sensory reception of such external speech stimuli.
INTEGRATOR
The second section of the system is called the integrator unit. Figure 1 shows that it is comprised of three basic components; the phase 1 integrator, the phase 2 integrator and the information storage component. Incoming information, in the form of speech sounds or other percepts, may be registered, retained, re called, or responded to by this unit. Phase 1 integration involves the recognizing and the attaching of significance to incoming stimuli; phase 2 integration involves the interpreting and the elaborating of incoming stimuli, in addition to the forming of verbal and nonverbal response attitudes. Information retention is subtended by the storage component which retains or releases stored information upon command.
Phase 1 integration represents the perceptualizing process served by the many primary sensory areas in the brain which recognize and pattern incoming auditory, visual, and tactile stimuli. Phase 2 integration represents the conceptualizing process served by the many secondary sensory areas of the brain which further process the various incoming stimuli in the manner already described.
Figure 1 also shows an error-measuring device existing within the unit. In the case of a speech response, this device compares the actual speech content with the prescribed speech content and determines the presence or absence of error performance. The concept of storage, or the recording of perceptual information by the brain, has been interestingly discussed by Penfield and Roberts (1959). They found that stimulation of certain portions of the temporal cortex during brain surgery on conscious patients resulted in vivid and complete re-experiencing of various past experiences by the patients. It would appear that something like a permanent registration of focused-upon sights and sounds takes place in certain temporal lobe brain mechanisms.
FIGURE 2. Anatomical schema of the speech system.
An example of the operation of the unit should contribute to the understanding of its function. In normal speech reception, the receptor unit receives acoustical as well as visual events associated with articulatory activity. Phase 1 integration consists of recognition of the acoustical and visual events as significant sound and sight stimuli which should be attended to, further processed, and possibly stored. The latter two steps are functions of phase 2 integration. For example, someone tells you the time is five o'clock. During this utterance, you recognize the spoken words as being pertinent auditory events and you attach meaning to them (phase 1 integration). Further processing of the utterance results in associations such as the utterance means: "It is time to go home"; or, "I cease working at that hour," and so forth (phase 2 integration). Storage of this information for future use may also occur. Additionally, the information represents a potential oral response, if, for example, someone should specifically ask you, "When do you stop working?" Figure 2 illustrates that perceptualizing and conceptualizing processes are carried out by both hemispheres. In the cybernetic analogue, all components are placed together within one unit to indicate the interrelatedness of all the processes, thus illustrating the constant interaction among perceptualizing and conceptualizing processes and information storage; such interaction tends to enhance and refine these three functions of the integrator unit. Also present in the analogue's integrator unit is the speech content comparator and speech content corrector, or the integrator unit's corrector device. Figure 2 shows this device in the temporoparietal region of the left hemisphere. The speech content corrector device functions as follows: Once the integrator has selected a certain response, it presents the neuronal pattern or nervous arrangement representing the idea to the cortico-thalamic area, which is the phase 2 transmitter component of the transmitter unit (the next unit to be discussed). This presentation automatically activates a neuronal pattern of corresponding words. In addition, the signal also has the potential for keeping the word neuronal-pattern active even after it discharges its pattern of signals into the phase 1 transmitter component, or the primary motor areas along the Rolandic fissure. Consequently, information in the form of speech is being sent out while at the same time word neuronal-patterns representative of the information being sent persist somewhere in the temporoparietal area in the left hemisphere. Via the auditory mechanism then, the trans mitted speech content is fed back to the temporoparietal area where actual speech content patterns are compared with intended speech content patterns. If discrepancies are found, or if changes appear desirable, there is scanning for different neuronal concept-patterns which, consequently, result in different neuronal word-patterns and hence different speech output.
TRANSMITTER
The transmitter section also has three basic parts and hence at least three functions. As already stated, ideas or speech intentions issuing from the integrator unit automatically excite word patterns in the phase 2 transmitter component which, in turn, activate appropriate signals in the phase 1 transmitter component.
Phase 1 transmission is responsible for exciting, simultaneously, the motor, generator, and modulator components of the effector unit which are actually responsible for producing the desired spoken words. In addition to these primary parts and functions, the unit also possesses a corrector device which operates as follows: The speech product comparator receives the input signals as well as the output feedback signals and determines the difference between the two; error signals, if present, represent the amount by which the command issued by phase 2 transmission has not been achieved by the effector unit. These error signals are then sent to the speech product corrector which combines error signal and input signal into a new corrected driving signal. The error signal also returns to the phase 2 transmitter component where it can trigger off the next command when the present output is error-free, or where it can hold the next command when the output contains error factors. This latter function represents a predictor potential existing within the speech product comparator ( Fairbanks, 1954) which allows command signals to flow rapidly without feedback monitoring when error-free sound products are anticipated. A similar predictor potential may be considered to exist within the integrator unit. The last component in the unit is called trans mission storage and represents the place where functional word patterns are stored. Another function of this component is activated when the individual be comes a listener and is receiving words; that is, words coming from a speaker activate corresponding word patterns in the listener's transmission storage section which, in turn, automatically excite associated ideas.
The phase 2 transmitter component may be considered to be the secondary motor speech area or cortico-thalamic unit whose cortical areas, according to Penfield and Roberts (1959), almost always are located in the left hemisphere.
Broca's area, the supplementary motor, and the temporoparietal areas are said to comprise the unit-the latter area is considered the most important. The phase 1 transmitter component represents the primary motor speech areas found along the anterior portion of the central fissure. These areas are responsible for innervating the respiratory-phonatory-articulatory muscle complex which produces the spoken word. As for the error-measuring function during word production, Ruch (1951) has made statements about the activity of the cerebellum which have a bearing on this component. In terms of guidance of movements, he conjectures that the cerebellum could be seen as the comparator component of a servo-mechanism. He indicates that it may receive signals from the cortex which represent the prescribed movement, and proprioceptive feedback signals from the muscles which represent the actual movement. Upon comparison or error measuring, if a discrepancy is found between prescribed and actual movements, appropriate error signals are then sent to the motor cortex which, in turn, alters its signals to the muscles and hence reduces the error.
In terms of function, let us suppose a 21-year-old individual is asked for his age. His speech reception-response mechanism proceeds as follows: (a) Recognition of the auditory events produced by the interrogator by the phase 1 integrator; subsequent interpretation by the phase 2 integrator; and scanning of the integrator's information storage component for the thought pattern corresponding to the idea of personal age. (b) Upon selection of the appropriate thought neuronal-pattern, there is an automatic activation of the appropriate word neuronal pattern in the phase 2 transmitter component. (c) The release of this word neuronal-pattern by the volitional mechanism excites the phase 1 transmission area along the Rolandic fissure which innervates the respiratory-phonatory-articulatory muscle complex needed to produce the word response, "twenty-one."
Figure 2 shows that the transmitter unit is comprised of the cortico-thalamic complex in one hemisphere, the primary motor speech areas represented in both hemispheres, and the comparator device represented, at least in part, by the cerebellum. In Figure 1, the analogue also shows an area for transmission storage which, as previously stated, represents the storage of word neuronal-patterns which the individual has developed and which are available to him.
EFFECTOR UNIT
The effector unit is directly responsible for the production of speech events.
It consists of three components: the motor, the generator, and the modulator.
[See Figure 1] The motor is responsible for producing the air column which supports speech, the generator is responsible for vibrating this air column or for voicing, and the modulator is responsible for breaking up the voiced air stream into particular articulatory units. The motor represents the respiratory structures; the generator represents the laryngeal structures; and the modulator the articulatory structures.
Figure 2 displays a larynx, tongue, and diaphragm to represent the effector unit. In Figure 1, the analogue shows all three components within one unit to indicate their interrelatedness.
SENSOR UNIT
The last section of the internal loop is the sensor unit. It has at least three components and is responsible for feeding back speech product and speech con tent data. Sensor 1 feeds back the auditory dimension of the sounds uttered; sensor 2 the tactile dimension; and sensor 3 the proprioceptive dimension. The unit may also include sensor 4 which represents the visual dimension; visual feedback would occur during mirror-speaking, for instance. Sensor 1 also feeds back the speech content.
To illustrate sensor unit functioning, let us suppose an individual has been asked for the name of his home town, which is Syracuse. However, also suppose that he has just left the dentist's office where he received some Novocain which is still in the process of wearing off, and thus he may be experiencing abnormal tactile and proprioceptive feedbacks. Under these circumstances, his response may be uttered as "Thyracuse." The sensor unit feeds back this speech signal to both product and content comparators. The content comparator will find the signal error-free since the name of the city is correct; however, the product comparator will find an error factor since the initial phoneme comprising the word product is incorrect. The product comparator then sends the error signal to the product corrector device for processing and this results in the immediate correction of "Thyracuse" to " Syracuse." Figure 1 shows the sensor components combined into one unit to indicate their interrelatedness. Two feedback signals may be seen arising from the unit, and they are the aforementioned speech product and speech content feedback signals. The sensor unit in Figure 2 is represented by pictures of: a muscle representing the proprioceptive end organs, lips in contact representing tactile end organs, and the ear representing the auditory system. When visual feedback is active, an eye may also be included here.
To summarize this section . . . , the speech system is made up of five basic units: the receptor, integrator, transmitter, effector, and sensor units. Both integrator and transmitter units include storage components as well as corrector de vices. The system has two outputs, namely, speech product and speech content.
SUMMARY
1. The speech system is described as a closed, multiple-loop system containing feedforward and feedback internal and external loops. Ten operations are de scribed as possibly taking place during a full cycle of speech behavior; these are: thought propagation, word formation, thought pattern-word pattern comparison, word production, actual word product-desired word product comparison, word product-thought pattern comparison, internal, multiple-loop speech recycling, word product-listener reaction comparison, actual listener reaction- desired listener reaction comparison, and internal, and external, multiple-loop speech recycling.
2. Descriptions of the various functional components of the speech system's internal loop are presented. Figures of a cybernetic analogue of the speech sys tem and an anatomical schema of the system are also presented. The speech sys tem is described as being made up of five basic units: the receptor, integrator, transmitter, effector, and sensor units.
REFERENCES
FAIRBANKS, G.: Systematic research in experimental phonetics: 1. A theory of the speech mechanism as a servosystem. J. Speech Hearing Dis., /9:133-139, 1954.
MEADER, C. L., and MUYSKENS, J . H.: Handbook of Biolinguistics, Part Two. Toledo, Weller, 1959.
MYSAK, E. D.: A servo model for speech therapy. J. Speech Hearing Dis., 24:144-149, 1959.
PENFIELD, W., and ROBERTS, L.: Speech and Brain-Mechanisms. Princeton, Princeton Univ. Press, 1959.
Rucif, T. C.: Motor systems. In: Handbook of Experimental Psychology, S. S. Stevens, ed., New York, Wiley, 1951.
VAN RIPER, C., and hiwiN, J . V.: Voice and Articulation. Englewood Cliffs, Prentice-Hall, 1958.
Also in Part 2:
- Communication Theory: Systems
- An Introduction to Cybernetics and Information Theory--Allan R. Broadhurst and Donald K. Darnell
- A Conceptual Model for Communications Research --Bruce H. Westley and Malcolm S. MacLean, Jr.
- A Transactional Model of Communication--Dean C. Barnlund
- A Helical Model of Communication--Frank E. X. Dance
- Speech System--Edward Mysak