|
This article is about an idea that has taken root during my 20 years of involvement with deafness at Queens University in Belfast. In a nutshell, the idea is that being involved with deafness calls for a particular kind of intellectual flexibility. That flexibility hinges on two things: 1) A willingness to recognize that your thinking could be dictated by a model of the human being that is not completely appropriate, and 2) A readiness to embrace different models if they perform a more useful function.
The idea runs against deep-rooted academic instincts. In a lot of areas, it is a good idea to settle for a particular model of the human being, then stick with it until you have extracted the last drop of useful insight. But, if you do that in the context of hearing loss, there is a high risk that you will become oblivious to the issues that actually matter. One aim of this article is to express the idea in a way that makes it easier for more people to understand and achieve the right kind of flexibility.
At Queens University, we began in the traditional academic fashion of taking very restricted models of the human being and applying them to specific problems (Editors note: For a background on models of the human being, see Glenys Chartrands three-part Concepts of Rehabilitation series in the Feb.-Apr. 2001 editions of HR.) In the process, we built up contacts with deaf and hearing-impaired people. This pushed us continuously to the conclusion that the models we were using at the time were acting as blindersallowing us to see some things clearly, but making it difficult to focus properly on other things that were obviously important.
We now believe that this progression reveals something fundamental about hearing loss. My hope in this article is that, by retracing our experiences, I can bring other people to understand some of these concepts in less than the 20 years it took me to grasp them.
Models of the Human Being
The phrase models of the human being has at its core an idea with a long history. Philosophers realized long ago that we cannot simply open our minds and let reality in; we only have access to an image of reality that our minds put together from incomplete evidence and their own innate resources. The roots of this idea go back to Plato, who lived about 400 BC. The Prussian philosopher Immanuel Kant made an important addition 2000 years later when he emphasized the role played by innate structures in our minds in shaping the way we put information together.
The 20th century saw several disciplines taking up that kind of idea. Thomas Kuhn1 linked it to the way science operates. He argued that scientists are guided by conceptual frameworks that direct them to look at particular issues, describe them in particular ways, and adopt particular approaches to explaining things. He called these frameworks paradigms. Similar ideas became important for work in the field of artificial intelligence. Scientists in that area argued that the way for a machine to make sense of incoming information was to have pre-structured packages to slot it intothese were called schemas, or frames, or scripts.2-4
Psychologists took up this theme and showed that human thought relied on similar kinds of pre-structuring.e.g., 5 They also applied the idea to their own discipline, and showed that differences within psychology were related to the use of different models of man.6 These models share the basic properties of other schemas or paradigms; they direct us to notice particular things, describe them in particular ways, expect certain kinds of patterns, consider particular kinds of action, and to blank out issues that the particular model does not handle.
There is no point lamenting the fact that people use models to understand each other and the world around them. Its part of being human. It is how we are, and it allows us to operate without being swamped by a flood of information and possibilities. However, using models to construct our individual realities exacts a toll. It means that we are always at risk of imposing a model that does not quite fit, or does not cover important aspects of some situations.
Questioning Simple Models of Speechreading & Speech Production
Since much of our research has been reported in detail in previous publications7-9 and would make little sense to duplicate here, this article will draw out the logical theme behind the detailsthe theme of the models we use to understand people, and the way these models affect the hearing care fields approaches to deafness.
Our research at Queens University began with a simple traditional model. For various reasons, we became interested in the problem of individual differences in speechreading. We were aware of other research that used the simplest possible model to describe these differences, placing speechreaders on a single scale running from no ability to flawless. The main question that we asked was whether one scale was enough to describe peoples speechreading ability.10 We concluded that it was not; there were several different skills involved, and people could be at quite different levels in the different skills.11
This seemed like a useful conclusion at the time, and it still does. If we know that people are weak in one aspect of a skill, then it makes sense to direct them either towards improving it or towards communication strategies that rely on it as little as possible. Our impression is that there is still resistance to taking even that small step away from the simplest possible model. It can be surprisingly hard to get past the image of a single line that progresses from good to bad.
Our second venture into deafness was prompted by another standard model that we were using in a different context. The model was a feedback loop: actions generate information, and that information is used to control what happens next. We wondered if and how speech changed when deafness cut out auditory feedback, and if this change could cast light on the feedback loop itself. The task seemed sensible because one of us had expertise in phonetic transcription, which provided the obvious way of analyzing the speech we recorded. The literature at the time suggested that deafness did not affect speech, but there was little phonetic analysis to support this hypothesis.e.g., 12
In fact, our phonetic transcriptions showed rather small numbers of errors.13 We might have stopped there if we had not spent time with the people we recorded. Because we had talked extensively with these people, we knew perfectly well that we had difficulty following their speech. We realized rather slowly that the problem was actually with phonetic transcription as a measurement instrument. Phonetic transcription was oriented towards certain kinds of problems, and those problems were not the most important part of what was actually happening.
So we moved away from transcription, and looked for instruments that would better describe what was going on. In the first instance, we measured how easily listeners could follow what deafened speakers said. Not surprisingly, they had problems, just as we had. Then we asked them about their impressions of the people they had heard. Again, not surprisingly, they formed various negative impressions of them on the basis of their speech.13,14
The problem reached a wide audience when Goehl & Kaufman15 claimed to have shown that post-lingual hearing loss did not affect speech production, and they concluded that recommendations for therapy were misguided. We answered that our evidence showed otherwise.16 The subsequent reply from Goehl & Kaufman17 illustrates the kinds of problems that a fixed model can create. They contended that the small number of transcription errors supported their case and, because anything else couldnt really be speech, it could be ignored. The argument is roughly the model used doesnt reveal a problem; therefore, there is no problem.
Researchers often get trapped into thinking that, if the effects cannot be described objectively, the effects cannot be real. And if the effects are not real (in terms of their description within the model), then they cannot be of any consequence. The next phase of our research was spent trying answer related questions in a constructive way. And to do that, we realized that we needed to think differently about the effects with which we were dealing.
Differences in Speech
In response to the question about reality, we moved into a purely acoustical analysis of deafened peoples speech. Our first serious attempt at that kind of analysis was reported in the 1992 book, Postlingually Acquired Deafness: Speech Deterioration and the Wider Consequences.7 We followed up this work by developing a program called ASSESS, which was used in a project coordinated by the Medical Research Council in England to evaluate cochlear implants.
Most of the research was reported in a 1995 paper18, showing that there are differences in the speech of deafened individuals. They tend to speak more slowly than controls, and, if anything, implantation makes this more pronounced. Their speech contains discontinuities, including silences and changes in the direction of pitch and intensity contours. Implantation improves some, but not all, of these discontinuities. Deafened people make use of volume patterns differently than normal-hearing people, and that volume pattern generally normalizes with implantation. Pitch range and movement are usually differentiated according to gender, but deafened people do not show usual pitch differentiation. Implants change these anomalous patterns, but not always in the right direction. The earlier study7 showed anomalies in the way energy is distributed across the spectrum. The later study18 had a glitch in that part of the program, so we were not able to confirm that pattern.
The above approach makes it quite clear that there are objective abnormalities in deafened peoples speech. In effect, programs like ASSESS embody a model that makes the effects apparent as oposed to concealing them. It is a large undertaking to do that. Nevertheless, it is only part of a genuinely satisfying approach. It needs to be linked with information about the functional significance of the variables that seem to be affected. Briefly, the speech anomalies seem to be involved in conveying information about the speaker rather than the words that he/she is producing. This idea has led us to a major project on how speech conveys information about the speaker, and a broad overview of that research has been published quite recently.19
What matters here is not the detail of the work on speech. It is the scale of the undertaking. We have been trying for 20 years to move from a general sense of the problems associated with deafened speech to a framework that has some chance of pinning down what the effects are and why they matter. We are still only part way. That is a useful pointer to the scale of the task that is involved in replacing restrictive models with more comprehensive ones that are genuinely satisfying.
What Do the Results Imply?
About the same time that we began developing acoustic analysis for speech, we began looking more systematically into the question of whether any of these speech effects actually mattered. We recognized that the question was part of a bigger, more fundamental issue. One of the most natural ways to model human beings is to assume that everyone has the same priorities. Of course, that is not necessarily true. Something that one person takes for granted may be an incredibly important objective for another (i.e., because it is easily available to the first person, and rare and elusive to the second).
We developed that point in collaboration with pyschologist Patricia Kerr.20 She carried out interviews with deafened people in the Belfast area, and asked them to identify aspects of their life experience that seemed important to them. On that basis, she developed a questionnaire that identified several recurring issues. She asked first how often these specific issues arose, and, second, how much impact they had on the quality of peoples lives.
As a first-step analysis, we took measures of peoples overall quality of life and state of mind, and compared two ways of predicting themone based on measures of the persons hearing and lipreading skills, and the other based on the subjective experience that they described. The experiential variables were far better predictors. In other words, the way people cope depends on their subjective world.
As a second step, we found six broad dimensions of variation in their subjective world. Most confirmed the importance of issues that we might have expected from the literature and our own contacts:
- Communicative deprivation: The sense of being denied access to the world of sound.
- Restriction: The sense of being unable to carry out activities that other people can.
- Malinteraction by hearing people: The sense of being subjected to hurtful behavior.
- Distress in interaction: The sense that their own behavior is not competent or normal
- Feelings of abandonment: An existential sense of being alone or isolated.
- Positive experience: The sense of having things to enjoy and take pride in.
The last one should not have come as a surprise to us, but it did. Being deaf does not mean being miserable, and one of the keys to a good life is finding things that you enjoy, whether you are deaf or not. The fact that happiness was missing from our model actually reflects a very widespread attitude. Thankfully, psychologists have begun to pay systematic attention to positive emotions, and they are currently a priority theme for the American Psychological Assn. Kerr has been developing the theme through work with Dai Stevens in Cardiff.
Restrictive vs. Inclusive Models
The progression that has been described up to this point can be expressed as shifts from a restrictive model of the person to an alternate model that is more inclusive. In terms of speechreading, we started from a model that assumes peoples competence can be placed on a single scale, and shifted to one where it takes several scales to specify where individuals belong. In terms of speech, we shifted from a model that regards people as producers of linguistic tokens to one that regards them as producers of a sound pattern capable of carrying several different kinds of information simultaneously. In terms of experience, we moved from a model implying that experience was bound to mirror objective reality (as we understood it), to one that accepted that subjective realities might differ enormously from person to personparticularly in the values placed on different aspects of objective reality.
The impetus behind those shifts was meeting deafened people when we recorded their speech. We were carried further in the same direction because the more recent research on experience meant more extended interactions with deaf and deafened people, and that gave us another push towards new kinds of modelsor rather two pushes.
The first push came from a deafened man named Frank Rice. Frank took an interest in our work on lipreading, and quickly came to the conclusion that the work missed all the interesting bits. What he did find interesting were models from cognitive psychology that described peoples information processing systems: extracting information from the signal, recovering it, combining it with stored information to draw new conclusions, and so on. He argued that cognitive models helped to make sense of his experience as a person who relied on speechreading, and he wrote a chapter on the idea that I warmly recommend. The title sums up his feelings about the literature that he had read including our papers, of course. He called it Do They Mean Us?21
Franks basic point was that there is no such thing as a free lunch. If you cant use the specialized resources of your auditory system to identify the words a speaker is producing, then you need to pull in resources from somewhere else. And if you use those resources for recognizing words, then you cant use them for other things such as working out what was meant but not said, or putting information into a form that can be remembered long term, or just doing something else while youre listening.
That idea sounds straightforward enough, but it has taken us a long time to come to grips with it. A series of students did experiments that suggest there is something to this idea, and their results were summarized in our 1992 book.7 Since then, a researcher named David Watson has been patiently developing the techniques to explore the idea properly, and at this stage, we are reasonably convinced that there is something there.
While it would be inappropriate to report details of Watsons work, broadly speaking it looks at memory for lists of simple items. The field of psychology has developed this topic in depth. There are well known effects in list learning that can be related to underlying mechanisms of memory, and if speechreading changes expected patterns, it provides strong evidence for ideas like those of Frank Rice.
Watsons research suggests that there are changes in two main areas. With normal-hearing people, the beginning of a list is usually better remembered than the middle. That changes in speechreading. The probable reason is that most speechreaders have to use all their mental resources right from the start. They dont have the luxury of having the spare capacity to organize material and file it away. The second area relates to cueing, (i.e., providing hints that help to organize material). Normally, people get benefit from cueing either before or after the main list. For speechreaders, only early cueing helps. Results like these confirm that relying on speechreading has effects on the entire way that people process informationnot just at the stage of registering words. The results also have some fairly obvious implications about ways of communicating with someone who relies on speechreading.
Models, Real World Practicality and Political Empowerment
David Watsons work comes from the push that Frank Rice gave us. As already indicated, we got a second push at about the same time when we got to know a lot of deafened people in Belfast. We gradually realized that they thought we should be making some practical contributions. We thought that was fair, since they were helping us with our research. So we helped with the practicalities of setting up a branch of the National Assn. of Deafened People (NADP). The point is that our involvement in this organization relates to the pattern of changing models. It meant switching from a model of deafened people as objects of study to a model of deafened people as a political entity.
The most fundamental aspect of that shift involved the issue of control. Deafened people felt that they were not in control of the things happening to them. Time and again, control was taken away from them because it was often inconvenient for a hearing person to share control with them. Quite often the people who took control went off in the wrong direction because they did not actually have the relevant information. And even if they did the right thing practically, it was offset by denying the deaf person the right to control his/her own life.
Organizations like NADP met a need because they first created a situation in which their members were in control, and second, gave members a chance to establish control over at least some aspects of life. None of that clicks unless your model of the person includes the fact that people need to feel in control of their own lives, or at least entitled to choose whether they take control or not. Your model should also include the fact that control is classically associated with access to information. Put those two together, and you see why it is dangerous to leave politics out of the picture when you are dealing with deafness.
That kind of shift was carried further in the 1990s when we became involved in two projects called SUCCEED and JUDE, which were designed to bring deaf and deafened people into Northern Irelands university system.22,23 As academics, we recognize that people are creatures molded by their educationthey crave education, they need it to function in society, and they are shaped largely by the education they receive and their experiences in the system. SUCCEED and JUDE forced us to realize how little we had understood this point. Signing deaf groups in particular cared passionately about education, and they put endless hours of work into establishing access courses that would bring their members into the universities. However, because their contact with academia had been so limited, they sometimes ran into problems because of their inexperience with academias ground rules. For example, one of the JUDE staff read that the university had been awarded a million pounds for research in chemistry, and wrote the university vice chancellor asking if he would consider giving it to the deaf unit instead. The point is that, if you dont know how universities work, this request seems perfectly reasonableand her previous experience had given her no way of guessing how universities work.
The last shift that I will consider arose from my own family. My father had a noticeable hearing loss when I was young, and his hearing loss is now profound. That brings me to one of the models that is most rarely talked about. It is the model of people as loved ones. My fathers deafness pains him because it cuts him off from family that he loves. It pains me because it cuts me off from the father I love. It pains me doubly because it changes him. When he has been cut off too long, he forgets how people communicate face to face, and this creates more barriers.
We tackle the barriers because walking away from them would mean walking away from love. The truth is surely that love is a large part of what makes deafness so difficult. Hearing loss would be easier to take if humans were constitutionally indifferent to other people. But love is actually one of the things that humans value most, and it is a disaster if we respond to difficulties by smothering it.
Talking about love tends to make people wince, so let me make it clear that it is not actually an eccentric thing to do. There are well-developed schemes for describing love and its effectse.g., 24 and theories of emotion increasingly recognize love as a basic part of the human makeup alongside fear, anger and other emotions.e.g., 25
How Academic Can You Be About Hearing Loss?
The point of tracing our research is to highlight choices that affect any serious attempt to think about people in general, and about deaf and deafened people in particular. To summarize the issues, Table 1 pulls together the options that have been discussed, puts them in a natural order, and adds an initial contrast that is obviously important, although it did not affect us directly. If nothing else, the table may be a convenient way of pointing out to people who would tend to gravitate towards the left-hand side that there are alternatives on the right-hand side that are worth at least a passing thought.
Reduced models: Often the most efficient ways to view problems | Fuller models: Often frustratingly messy, but sometimes needed |
Biological systems | Psychological systems |
Located on a single scale of competence | Located in a multidimensional space of competencies |
Engaged in detecting signals | Engaged in high-level information processing |
Engaged in producing words | Engaged in speech communication that affects listeners at multiple levels |
Inhabiting objective situations that can be observed by anyone | Inhabiting worlds that are subjective, particularly in matters of value |
Victims of problems, recipients of support | Concerned with achieving control through political means |
Recipients of knowledge | Engaged with educational systems, or estranged from them |
Self-contained individuals | Engaged in loving relationships, or estranged from them |
Table 1. A review of the different models discussed in this article. |
The easy way to end the paper would be to pronounce that the left-hand side represents old-fashioned misconceptions, and the right hand side represents the way we should be thinking. That would be easy, but unfortunately, it would also be dishonest. Trying to reach an honest conclusion leads back to the question in the title of the paper.
The truth is that the table describes the problem we face. It describes two orientations. We cannot do without either, and yet they are monstrously difficult to bring together. We know because we have spent 20 years trying.
The right hand side of the table is not our invention. Ideas like that have been raised repeatedly by people who are personally involved with deafness, and some of the most persuasive are published self-reports by deaf and deafened people themselves. They have written about their own experiences and their involvement with political and educational causes, often very eloquently.e.g., 26-30 The problem is that somehow that literature stays on the fringes of societys consciousness.
The left-hand side corresponds much more closely to the views that dominate academics and the professions with a strong academic orientation. The framework that has been used here suggests a way of summarizing how we in academia tend to operate. Part of the ethos is seeing other people through the filter of models that are quite restrictive, and agreeing to behave as if the things we filter out do not exist. For instance, conference speakers who talk about such doubtful things as love are very liked to be hauled up for unprofessional emotionalism.
That ethos can be deeply frustrating at times, and yet the truth is that it has great strengths. As I pointed out earlier, it is closely linked to what Kuhn identified as a central feature of science. Science and other academic enterprises work because the people in them agree to operate within a paradigmwhich means shutting out questions that the paradigm assumes are either answered or unanswerable. Outright rejection of an instrument as powerful as that is not a sensible option.
In a nutshell, the problem is that people need simplified models to keep their thinking straight and clear, and complicated models to do justice to reality. If our group has a claim to be distinctive, it is not because we recognize the relevance of complicated models. It is because we have tried to find ways of moving towards them without letting go of the ideals that pull people towards simple models. That means looking for frameworks that have the academic virtues of clarity, logic and simplicity, as well as the complexity to do justice to reality.
To answer the question, How academic can you be about deafness? we can be academic about issues on the left side of the table, but only partly about issues on the right sideat least, not academic as we understand it. Pushing the boundaries will require a huge amount of work.
One aim of this paper is to encourage people to take on that work. The other is to suggest the kind of spirit that it makes sense to adopt until the work is done. To come back to the beginning, the key word is flexibilitymoving between imperfect models in a way that recognizes both their value and their limitations. My hope is that this may be a little easier if we understand that our problems come from basic features of the way that humans understand reality and each other. It is human nature that we can only see each other through the filters of our models.
The remarkable thing is that we can learn to understand the limitations of our models, and move intelligently between them. That is an ability that still sets human beings apart from computers that have artificial intelligence. It seems to me that we should recognize the value of this ability, and make use of it.
Roddy Cowie, PhD, (r, pictured with Geoff Plant) is a psychologist in the Dept. of Psychology at Queens Univ. in Belfast, Northern Ireland. |
Correspondence can be addressed to HR or Roddy Cowie, D Phil, School of Psychology, Queens University, Belfast BT7 1NN, Northern Ireland, United Kingdom; email: [email protected]. If you would like to obtain a copy of the conference proceedings, or find out more about the next conference (to be held in May 2003), contact: Geoff Plant, The Hearing Rehabilitation Foundation, 35 Medford Street, Somerville, MA 02143; email: [email protected]; Web site www.hearf.org.
References
1. Kuhn T: The Structure of Scientific Revolutions. Chicago: Univ of Chicago Press, 1962.
2. Rumelhart DE: Notes on a schema for stories. In DG Brobrow & AM Collins (eds) Representation and Understanding. New York: Academic Press, 1975.
3. Minsky M: A framework for representing knowledge. In P H Winstons (ed) The Psychology of Computer Vision. New York: McGraw Hill, 1975: 211-217.
4. Schank RC & Abelson RP: Scripts, Plans, Goals, and Understanding. Hillsdale, NJ: Erlbaum, 1977.
5. Anderson RC & Pichert J: Recall of previously unrecallable information following a shift in perspective. J Verbal Learning and Verbal Behav 1978; 17: 1-12.
6. Chapman AJ & Jones DM: Models of Man. Leicester, England: British Psychological Society, 1980.
7. Cowie R & Douglas-Cowie EE: Postlingually acquired deafness: Speech deterioration and the wider consequences. Berlin: Mouton de Gruyter, 1992.
8. Cowie R & Douglas-Cowie EE: Speakers and hearers are people: Reflections on speech deterioration as a consequence on acquired deafness. In G Plant & K-E Spens (eds.) Profound Deafness and Speech Communication. London: Whurr Publications, 1995: 510-527.
9. Cowie, R, Watson D, Kerr P, & Douglas-Cowie EE: Psychology and hearing impairment: Focussing on the people with the loss. Irish Jour of Psychol 1995; 16: 288-298.
10. Cowie R, Gailey EL, McCorry M-L & Burns R: Speechreading: structure and variation in a skill. J Northern Ireland Speech and Language Forum 1982; 8: 57-69.
11. Gailey EL: Psychological parameters of lip-reading skill. In B Dodd & R Campbells (eds) Hearing by Eye: The Psychology of Lip-Reading. London: Lawrence Erlbaum, 1987: 115-141.
12. Bower T: Human Development. San Francisco: W.H. Freeman, 1979.
13. Cowie R. & Douglas-Cowie EE: Speech production in profound postlingual deafness. In M Lutman & M Haggard s (eds.) Hearing Science and Hearing Disorders. London: Academic Press, 1983: 183231.
14. Cowie R, Douglas-Cowie EE & Kerr A: A study of speech deterioration in post-lingually deafened adults. J Laryngology and Otology 1982; 96: 101-112.
15. Goehl H & Kaufman D: Do the effects of adventitious deafness include disordered speech? J Speech Hear Dis 1984; 49: 58-64.
16. Cowie R., Douglas-Cowie E & Stewart P: A response to Goehl and Kaufman (1984). J Speech Hear Disorders 1985; 51: 183-185.
17. Goehl H & Kaufman D: The real thing: A reply to Cowie, Douglas-Cowie and Stewart. J Speech and Hear Disorders 1986; 51: 185187.
18. Cowie R, Sawey M & Douglas-Cowie EE: A new speech analysis system: ASSESS (Automatic Statistical Summary of Elementary Speech Structures). In K Elnius & P Branderuds (eds) Proceedings of the XIIIth International Congress of Phonetic Sciences, Stockholm (Vol. 3), 1995: 278-381.
19. Cowie R, Douglas-Cowie EE, Tsapatsoulis N, Votsis G, Kollias S, Fellenz W & Taylor J: Emotion recognition in human-computer interaction. IEEE Signal Processing Magazine 2001; 18 (1): 32-80
20. Kerr P & Cowie R: Acquired Deafness: a multidimensional experience. Brit Jour of Audiol 1997; 31: 177-188.
21. Rice F & Cowie R: Do they mean us? Preliminaries to a theory of lipreading that is recognisable to deafened lipreaders. Belfast Working Papers in Linguistics 1990; 10: 17-28.
22. Cowie R & Wylie J: Opening Higher Education to People with Impaired Hearing. Belfast: RNID, 1997.
23. Wylie C & Cowie R: Frameworks for Development: Deaf and Partially Hearing People Growing Up in Northern Ireland. RNID: London, 1997.
24. Sternberg R: A triangular theory of love. Psychological Review 1986; 93: 119-135.
25. Oatley K & Jenkins J: Understanding Emotions. Oxford: Blackwell, 1996.
26. Abrahams P: The effects of deafness. Hearing 1972; 27 (10): 295-297
27. Ashley J: A personal account. In H Orlans (ed) Adjustment to Acquired Hearing Loss. London: Taylor and Francis, 1985: 59-70.
28. Savil S: Acquired hearing loss: A personal view. Deafness 1988; 4 (1): 6-7.
29. Wood P: The world we have lost. In J Kyles (ed) Adjustment to Acquired Hearing Loss: Analysis, Change, and Learning. Bristol, England: Centre for Deaf Studies, Univ of Bristol, 1987: 156-162.
30. Woolley M: Acquired deafness: acquired oppression? In J Kyles (ed) Adjustment to Acquired Hearing Loss: Analysis, Change, and Learning. Bristol, England: Centre for Deaf Studies, Univ. of Bristol, 1987: 169-175