Interpretation, Truth and Correspondence

Transcription

1 Journal for the Theory of Social Behaviour 30: Interpretation, Truth and Correspondence 377 Interpretation, Truth and Correspondence PENTTI MOILANEN In reviews of qualitative research one can sometimes see complaints about inadequate justification for the methods, results and conclusions of qualitative research. There are strong reasons for scepticism when reading research reports (see Howe & Eisenhardt 1990, 2; Miles & Hubermann 1994, 2; Silverman 1994, 153). Shortcomings in the evaluation of validity are at least partly due to vague conceptions of the nature of validity or truth of qualitative research (Howe & Eisenhart 1990, 2). The only way to overcome these shortcomings is to engage in philosophical investigation of the modes of truth in qualitative research. In this analysis it is not useful to seek support in notions like positivism or post-positivism or naive empiricism because these notions are too vague for serious analysis. Instead one should have a closer look at the aims and methods of interpretation. Because all knowledge is based on presuppositions and is a social product, the starting points and aims of research are not value neutral (see Altheide & Johnson 1994, 488). In evaluating the validity of qualitative research it is possible to assess the starting points or aims of research beside its truth. Therefore truth of interpretations is only one criterion for the quality of qualitative research, but it is a crucial one. The correspondence theory of truth is often considered unsuitable for qualitative research. One idea behind the rejection of the correspondence theory of truth is the conception that one cannot test the correspondence of interpretations and social reality because social reality consists of changing interpretations. According to this conception there is no ultimate social reality with which the interpretations made by a researcher could be compared. From the constructionist viewpoint people are constantly constructing their psychic and social realities shaped by historically contingent conventions of discourse (Gergen 1985, 272). I think that relying on the correspondence theory of truth is still possible. One aim of social research is to reconstruct the interpretations made by the members of the societies investigated. Here one can, in principle at least, compare the. Published by Blackwell Publishers, 108 Cowley Road, Oxford, OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA.

2 378 Pentti Moilanen primary interpretations under investigation with the interpretations made of these primary interpretations. Strauss and Corbin (1990, 252) see three kinds of judgments about the validity, reliability and credibility of data. Judgments are made about the data, research process and empirical grounding of the research findings. In research practice there seems to be an emphasis on the validity of the research process (Atkinson, Health and Chenail 1991, 162). It is supposed that the research process should be made as transparent as possible to allow the evaluative analysis of each step. Here I shall take another course of analysis and concentrate on the third form of validity analysis. The starting point of my argument is that one tries to find some independent support for an interpretation hypothesis. This support is based on a comparison of the interpretation with certain facts. Validity in qualitative research 1. TRUTH AND INTERPRETATIONS Truth is the most important notion of science. One of the aims of science is to build up a truthful conception of reality. The requirement of truthfulness is a demanding one because there is no direct way to evaluate the truthfulness of scientific theories. In evaluating theories one is relying on other theories (see Longino 1979, 39 40). This notion leads us to question the usefulness of the notion of truth in science. Should we have some easier ways of evaluating the quality of scientific theories? Maybe it is enough that theories are useful in the explanation of phenomena or in the prediction of courses of events or in coping with phenomena in our daily lives. The problem of truth has become a complex question in qualitative research. The methods of gathering and interpreting data are not simple procedures that could be used in exactly the same way by every researcher. Some researchers are more sensitive than others in interviewing people or more insightful in construing interpretations. The process of interpretation is therefore a subjective process. Especially in action research the subject of research changes due to the research process. It may be argued that in research where the methods of data collecting cause reflection in the people studied, the research cannot objectively describe social reality. Different research procedures change people in different ways and therefore there is no objective way to compare the results of various investigations. A third problem concerns the ontology of social reality. This problem is tied to the subjective nature of social reality. Social reality consists to a large extent of interpretations (Lenk 1991, 291). Many social facts are based on interpretations that are not alike in different societies. For example democracy, equality or freedom have no universal meaning. In these cases the researcher makes subjective interpretations of subjective interpretations. Is there any place for the notion of truth?

3 Interpretation, Truth and Correspondence 379 In qualitative research the traditional notions of validity and reliability are seldom used. Instead one uses notions like credibility, transferability, dependability and corfirmability (Lincoln & Guba 1985, 300). These notions better suit the subjective nature of the research process. For me the most interesting of these concepts is the notion of credibility. It seems to presuppose the notion of truth. The other concepts are coherent with the subjective side of interpretations but the notion of credibility is different. It requires that one makes interpretations that correspond to reality, even if this reality is a subjective one. Subjectivity and objectivity of interpretations The reality investigated is in many cases subjective and the process of interpretation is subjective, too. This does not, however, mean that the reality and the process of interpretation are alike because the term subjective has different meanings. Searle (1995, 8) writes that we have to distinguish an epistemological and an ontological use of this term. In the epistemological sense subjective and objective refer to the evaluation of knowledge or opinions. In this sense musical taste is subjective because there are no objective criteria for music. One person prefers rock music and someone else jazz. It depends on one s own musical taste what kind of music is considered good. Objective facts do not depend on people s opinions or feelings. Helsinki is the capital of Finland and this is an objective fact. In the ontological sense objective and subjective refer to the way of being. Some entities exist in private experience. Pains are subjective in this sense. Their existence presupposes some person to experience the pain. In this sense it is possible, at least in principle, to have objective knowledge of subjective facts. When I say that you have a headache, I may make a correct or incorrect statement about your experience. Even if I cannot objectively say that jazz is better than rock music, I still can say that you prefer jazz. Even if the social world consisted only of subjective interpretations, it is not a mistake to state that we can, in principle at least, make objective interpretations of social reality. Even if these interpretations are not the whole truth, they can accurately describe the subjective social world from some point of view. In this sense we can have various true interpretations of the social world, which consists only of subjective interpretations. I think that two specifications must be made here. First, the social world does not consist only of subjective interpretations. Secondly, even if objective interpretations of social reality are in principle possible, it does not have to be so in practice. Lincoln and Guba (1985, ) state: When naive realism is replaced by the assumption of multiple constructed realities, there is no ultimate benchmark to which one can turn for justification... In order to demonstrate truth value, the naturalist must show that he or she has represented those multiple constructions adequately, that is, that the reconstructions... that have been arrived at via the

4 380 Pentti Moilanen inquiry are credible to the constructors of the original multiple realities. To my mind this idea is consistent with the correspondence theory of truth and fits in with the two meanings of subjectivity. Coherence and correspondence in hermeneutics The correspondence theory of truth is not usually considered to fit in with hermeneutics (see Wachterhauser 1994, 24). There are at least three reasons for this. Hermeneutics has construed the interpretation process as a dialogue based on a hermeneutic circle. Interpretation has not been seen as a hypotheticdeductive procedure including testing of a hypothesis, but as a gradual correction of pre-understandings. Secondly, one of the aims of hermeneutic understanding is mutual understanding that leads to edification. Here, truth means truth to a person, not universal truth. Third, the coherence theory of truth fits in with the actual methods of hermeneutic interpretations. In interpreting literary works, one aims at building a coherent interpretation of the work. In searching for hidden meanings or presuppositions one cannot compare the interpretation hypothesis with a certain fact given in the text because the hidden meanings and presuppositions are not visible on the pages. The existence of hidden meanings and presuppositions is non-visible. When one wants to understand the action of certain people, one has to try to see the world through their eyes. Even this is not enough because their action is situated in some culture and in interpreting individuals one must make interpretations of the cultural ethos of their society. In interpreting text or other social facts one is trying to construe an interpretation that makes the subject of interpretation a meaningful totality. Interpreting social facts is woven into the interpretation of culture. The interpreter does not concentrate on separate facts but on the entire culture. The idea of coherence theory of truth fits in with this endeavour. 2. VERIFICATION OF INTERPRETATIONS There are many kinds of qualitative research and even of interpretive research. For instance Miles and Huberman (1994, 7) list more than twenty types of qualitative research. Therefore it is not possible to deal with the truth of qualitative research in its totality but with the problems of truth in some certain kind of qualitative research. In this article I deal with the interpretation of the reasons for action. In interpreting the action of some people the researcher may try to find the reasons for their action. These reasons are needed to make the action, previously regarded as odd, seem meaningful. If the interpreter knows what the agent aimed at with his action and why he considered that aim valuable, the interpreter can understand the action of the agent. In short, one is making

5 Interpretation, Truth and Correspondence 381 interpretations of people s intentions and beliefs in order to understand their action. (Carlgren & Lindblad 1991, 5009; Nespor 1987, 323; Noel 1993, 142.) Understanding and explanation of action are tied together. Reasons contribute causally to action and therefore the intentional explanation of action can be seen as one type of causal explanation. On the other hand knowledge of the reasons for action makes it possible to understand the action, i.e. to see its point and to re-identify it. The agent s conceptions of the reasons for his action may sometimes be misguided. Therefore to see the agent s point of view may help to understand the action from his position but it does not necessarily give a valid explanation of the action. In these cases the interpretation is one-sided and it needs correction. Still, in some sense, these kinds of invalid explanations provide some understanding of the action studied. One can see what the agent considers his action to be like. Giving a correct intentional explanation of action does not guarantee its understanding. The interpreter may lack acquaintance with the culture the person under investigation is living in and therefore the interpreter is unable to see the cultural dimensions of factors affecting action (see Martin 1976, ). A correct understanding of an action is linked to a causal role of the reasons. As Bhaskar (1979, 115) writes: It is only if X does A because of R that we are justified in citing R as the reason for A t. And there would seem to be no way of explicating the because save in terms of causality. For unless the reason was, in context, a necessary condition for the physical movement that actually occurred... then... the particular reason explanation given must appear as totally arbitrary. Like a good fairy-tale it may soothe and satisfy the listener, but it cannot explain. If and whenever they explain, then, reasons must be interpreted as causes, on pain of ceasing to explain at all. (Bhaskar 1979, 115) According to agent causalism the reasons as mental states do not in themselves cause the action but the agent is this cause. The reasons causally contribute to action because they have a part in the control of action (see Vollmer 1995, 184). This control is partly conscious and partly unconscious. It is possible that the same action can be explained in different ways. We humans are individuals capable of conscious control of our action, holders of social myths, reproducers of social processes and animals with various animal instincts. An explanation of action may refer to any of these phenomena. Therefore the reasons explanation is only one possible kind of explanation. In some cases these other kinds of explanations are explanations for having certain reasons for action and therefore they complement the reasons explanations. Sometimes they are rival explanations that aim at making reasons explanations unnecessary. Even if we are convinced of the correctness of the reasons explanation it is always possible that there are other or better explanations to be sought after. That is why the correctness or validity of a reasons explanation is relative to this special perspective or language game. Therefore it is presupposed in my argumentation

6 382 Pentti Moilanen that reasons explanation is a valid type of explanation even if it is only one possible way to explain the action investigated. The reasons for action do not constitute a uniform case. There are different kinds of action and different kinds of reasons for action. Besides deliberated action with conscious reasons there is routine action. In the cases of routine action, the action may have been deliberated upon first but gradually become routine. Another possibility is that the action has never been deliberated. In any case the reasons for action are tied into the control of action by the agent. The agent (consciously or unconsciously) aims at something and alters (consciously or unconsciously) his action according to the changing situation in order to attain the goal. The control of action presupposes what Searle (1992, ) calls the background of action. This background includes, e.g. habits that are automatically realized in action. For example, when walking into a restaurant, we do not have to wonder how to stay upright while walking. It succeeds automatically. In this way the reasons for action are tied to a conscious and unconscious control of action and background capacities. This holds also for non-physical action. Even if there are many kinds of action and many kinds of reasons for action the agent always controls his action somehow. It will be seen later in this article that this notion is very important when discussing the evaluation of the truth of interpretations concerning the reasons for action. To evaluate the usefulness of the coherence and correspondence theory of truth for qualitative research one must have some concrete examples. Both these theories are in principle useful, but usefulness is not a matter of principle but of concrete cases. Therefore, I use a short interpretation example and show various ways to evaluate these interpretations. After this I discuss the usefulness of the two theories of truth. Let us suppose that the researcher is observing and interviewing a teacher and tries to find out what he aims at in teaching his pupils. His teaching sometimes seems progressive and sometimes very traditional, but can be labelled as neither. After various interviews and observations the interpreter has come to the conclusion that the teacher is mainly interested in the development of the thinking skills of his pupils, but his conceptions of thinking skills are very vague. Therefore, his action is inconsistent and even ineffective. How can the interpreter evaluate these interpretations of the reasons for action? I think that there are at least four different procedures for evaluating these kinds of interpretation: participant agreement (member checks), consensus of interpreter community, prediction of action and coherence of the interpretation. I will now examine each of these procedures in turn. Participant agreement Participant agreement is quite often used in qualitative research (see Ashworth 1993 and Terhardt 1985 for the principles of participant agreement). It is supposed

7 Interpretation, Truth and Correspondence 383 that the person being interpreted has a special relation to the reasons of action. It is his action and the reasons for his action that are the subject of interpretation. Here the interpreter has to ask the agent if the proposed interpretation is true or false. If the reasons for action were subjective and conscious in their ontological status, this method of verification would be valid. Unfortunately the reasons for action are not equivalent to the agent s conceptions of the reasons for his action. He may have an incorrect conception of the reasons because he does not want to see the negative sides of his action. There are also implicit reasons especially in routine action. Besides, there are implicit presuppositions about reasons for action, of which the agent is unaware. Sometimes the agent has planned his action carefully. He has considered various aims and various methods of apprehending these aims. In these cases, the reasons for action are the mental states involved in the planning of action. Therefore it is possible to think that the evaluation of truth of interpretation is based on a comparison of the interpretation and the mental states involved in the planning of action. It is possible to comprehend the truth of the interpretations made of people s experiences in the same way. Truth is based on the comparison of interpretations and experiences. In these cases the idea of the correspondence theory of truth is well-founded. The interpretations of implicit reasons for action or reasons for routine action cannot be compared with conscious mental states that control action because this control is not conscious. In spite of this, the idea of participant agreement works. It may be supposed that the agent has a special relationship to his unconscious beliefs or intentions. He can recognize them because previously he has been aware of them. In these cases the interpreter compares his interpretations with unconscious mental states of the agent. Sometimes the agent can remember the beliefs or intentions that previously were conscious but are now forgotten. Sometimes he cannot remember them because he controlled his action without being conscious of them. In both cases the idea of the truth as a relationship of interpretation and psychic reality is a sound one, even if the actual comparison were not feasible. Unfortunately there may be some limits here because all implicit reasons of action cannot be conceived in this way. Nevertheless, with the aid of the idea of unconscious control of action, we can rescue the idea of a comparison of the interpretations and subjective reality. However, the validity of the method of participant agreement is questionable here because the agent does not have direct access to this unconscious reality. The presuppositions of action or of the reasons for action are usually not mental states. For example, the teacher presupposes that his pupils are living humans and can think but he has never had these kinds of conscious thoughts. In the same way I have opened the door and never wondered if the handle would bite my hand. I just presuppose this in opening the door (see Searle 1992, 185). The agent does not have direct access to this kind of presuppositions. For this reason

8 384 Pentti Moilanen participant agreement is not suitable for the interpretation of the presuppositions of action or of the reasons for action. It seems that the correspondence theory of truth is meaningful with participant agreement, if one is interpreting the reasons for conscious action. In the case of routine action the method of participant agreement seems invalid even if the correspondence theory of truth held in these cases. In the case of interpreting the presuppositions of the reasons for action the correspondence theory does not seem meaningful. This is so because these presuppositions are neither conscious nor unconscious mental states. They belong to the non-intentional background of human action (see Searle 1992, 195). Communal validation There are several methods of validation that stress the importance of research collectives in the evaluation of the quality of research: peer debriefing, auditing and using inter-judge reliability. The common idea behind these methods is that to detect the bias and to evaluate the quality of research there is a need for other evaluators besides the researcher himself. According to Atkinson, Health and Chenail (1991, 163), the trustworthiness of hypotheses cannot be established by an individual researcher. Legitimization of knowledge requires a judgment of a community of the consumers of the research and it must be a democratic process. Communal legitimization presupposes an unforced dialogue where everyone can state his arguments pro and contra a hypothesis. Auditing is a process for reviewing, analysing, and assessing research documents (Whitmore & Ray 1989, 79). Among other things the audit seeks to decide whether interpretations are grounded in the data, and whether inferences are logical. However, the audit does not re-analyse the data. In peer debriefing a second investigator is cooperating in the whole research process (Lincoln & Guba 1985, 308). He functions as a dialogue partner, for example to test working hypotheses. Inter-judge assessment is used for instance in phenomenographic research to evaluate the category systems developed (Sandberg 1997, 206). This idea of communal validation is founded on the consensus theory of truth developed by Habermas. Habermas (1973, 215) criticizes the correspondence theory of truth for its false conception of facts. To his mind facts depend logically on the theories they are presumed to verify. There are no theory-neutral facts that theories could be compared with. The validation presupposes instead a community of researchers that enter into an unforced dialogue where the participants can evaluate the strength of various arguments. I think that the consensus theory of truth misses an important point. Communicative validation by itself is not enough to validate some hypotheses or interpretations. Agreement, even unforced agreement, is not the most crucial thing.

9 Interpretation, Truth and Correspondence 385 It is not enough that investigators share a view but the consensus has to be based on valid arguments. The arguments have to be convincing. These arguments point to somewhere outside the dialogue. There outside the dialogue we may find the real criteria of truth used in communicative validation. These criteria fit in with coherence or correspondence theory of truth, I suppose. Predictions Our everyday life is based on the possibility of predicting people s action. People act according to some rules and according to their beliefs and intentions. When we know these rules or beliefs and intentions, we can predict the action of others in various situations. Therefore, it is also possible to use predictions to verify the interpretations concerning the reasons for the action of other people. These predictions may involve active or passive participation from the researcher. He may try to change the agent s situation or he may just wait and see what happens. The active way is of course more effective. In the example case the researcher wants to change the situation of action. He does not have to make any dramatic changes, but uses the fact that he himself belongs to the agent s situation. The researcher may make changes in this part of the agent s situation. (See Gruschka & Geisler (1982, 632) for another example.) If it is true that the agent is interested in the thinking skills of his pupils, he should be interested in various theories concerning these skills and in various practical methods of developing them. The researcher can give him some hints on these theories and methods. If the teacher wants to know more about them and use them, the interpretation hypothesis is at least partly confirmed. Of course the verification procedure is not this simple because it is based on various presuppositions concerning the personality of the teacher, his life situation, the culture of the school and so on. For instance, if the teacher does not appreciate researchers, he does not want to learn anything from them, and does not show any interest in these ideas for developing thinking skills. In spite of all these reservations, the procedure described here has many affinities with the experimental method. It is supposed that knowing the causes of phenomena makes it possible to predict changes in these phenomena. This idea fits in well with the correspondence theory of truth. We can compare our beliefs with reality by the experimental method. The prediction of human action is problematic. According to agent causalism intentions and beliefs do not cause action. The cause of action is the agent. It is always possible that the agent changes his mind. Secondly, the situations of human action are very complex. Considering all the relevant factors in making forecasts about the action is quite difficult. Third, there may occur some surprises that change the action situation. Finally, acting in different ways for the same reasons is possible. (See Taylor 1971, 49.)

10 386 Pentti Moilanen I think that predicting people s action is quite simple, but to analyse what the actual outcomes mean is more complicated. Even when predictions fail it is possible that an interpretation hypothesis is verified because other phenomena may be the cause of the failure. Even when predictions are realized, the interpretation hypotheses do not necessarily become confirmed. In evaluating the validity of the interpretation hypothesis one cannot rely on individual facts but on a whole group of facts joined in complex ways. Therefore, the interpreter is constructing a large scale picture of the agent, his situation and his action. He is trying to construct a coherent large scale interpretation. Coherence Evaluating the coherence of an interpretation hypothesis is the fourth procedure I mentioned. In this method it is presupposed that people are trying to live a coherent life. The values, beliefs and action of people fit nicely together. The coherence of life is important because it makes everyday life easier. Therefore, it is justified to construe coherent interpretations of people s action. If what people say and what they do are not in accordance it is not reasonable to believe in what people say. Or if their utterance is inconsistent, suspicion that they are lying is reasonable. In the same way the interpreter must search for inconsistencies or discrepancies in his interpretations. Quite small hints may lead him to suspect the interpretations. In this fourth case of verification of interpretations the role of the correspondence theory of truth is not obvious. The correspondence theory of truth presupposes some kind of comparison of reality with interpretations. It seems that in analysing the coherence of interpretations one is concentrating wholly on the interpretation and forgets the reality outside the interpretations. At a closer look we can notice that in analysing the coherence the focus of the interpreter s attention is not focused wholly on his interpretations. He is also observing what people do, how they spend their time and so on. Why is this observing necessary? I think because the interpreter wants to know what the real reasons (e.g. reasons that causally contribute to action) are. He is not interested in the opinions of the agent, in his stories, but in the reasons that have this causal role. Therefore it is not enough to construe a coherent interpretation, but an interpretation is needed that makes it possible to explain the action. In this way the idea of comparison of interpretation of the reasons for action with the action itself is hidden in the evaluation of the coherence of interpretations. Nevertheless, this comparison means something else than the idea of the correspondence theory of truth. The idea of coherence in this case is that action and the interpretations of its reasons must fit together, not that the interpretation is compared with something assumed to be the real reason for that action.

11 Interpretation, Truth and Correspondence INTERPRETIVE AWARENESS In analysing phenomenography Sandberg (1997, 207) argues that researchers cannot escape their subjectivity in research because the results are always constituted through the researcher s interpretation. For that reason the use of interjudge reliability leads to contradiction with subjectivist epistemology. According to Sandberg this procedure is based on objectivistic epistemology. Instead Sandberg (1997, 209) suggests using interpretative awareness as a method for estimating the reliability of his results: That is, in order to be as faithful as possible to the individuals conceptions of reality, the researcher must demonstrate how he/she has controlled and checked his/her interpretations throughout the research process... It is important to notice how the interpretations and preconceptions of the researcher influence the research process. Sandberg (1997, 209) denies the importance of the correspondence theory of truth for phenomenography: So, based on the epistemology of intentionality, establishing reliability of the researcher s interpretation is crucial; reliability of results relating to objective reality falls outside the domain of interest. At first sight this argument seems coherent. To be as faithful as possible to the conceptions of the people investigated, one must be alert to bias. The researcher must withhold theories and prejudices when he interprets an individuals conceptions. This is the most basic idea of phenomenology. To my mind the most crucial question is why it is important to be as faithful as possible to the conceptions of people investigated. The answer is simple. Because we want to find out how they think or conceive their world in reality. This ontology presupposes that there are people who think and the interpretations of their thinking must be as truthful as possible. This means that the correspondence theory of truth is detectable here. The weakness of Sandberg s argument rests to my mind in the dichotomy of subjective and objective epistemology. In epistemological analysis we have to consider both sides, the preconceptions woven into the interpretation process and the subject of interpretation, and the dialogue between these two. Therefore, interpretive awareness cannot be the foundation of the validity or trustworthiness of research. Like communicative validation it misses what it means to be a true statement of subjective reality. 4. TRUTH OF INTERPRETATIONS Theories of truth cannot be verified. Therefore, these theories cannot prove anything about qualitative research. In spite of this they are needed to support qualitative research. They are like stories that give meaning to the efforts of researchers (see Koski 1996, 13). These stories can tell why it is not necessary to reach for absolute verification, or why there is disagreement among researchers.

12 388 Pentti Moilanen Another reason for their importance is that they make it possible to ask different kinds of questions about the results of qualitative research. Not all theories of truth make it meaningful to ask, for example, whether there is a valid consensus in a research community. It is impossible to prove that the correspondence theory of truth is true. This theory is a philosophical theory and therefore the question of its truth is not meaningful. The more fruitful question is what meaning this theory has for research activities. I think that a transcendental argument is sound: the correspondence theory of truth is presupposed in the many activities of researchers who use qualitative methodology. (See Bhaskar 1979, 7.) Now one can see why the idea of the correspondence of theories and reality is still sound. One aim of science, even of hermeneutics or phenomenology, is to make correct statements or interpretations of reality. This aim makes the correspondence theory of truth useful. The correspondence theory of truth is important in qualitative research because it makes is possible to ask which of the competing coherent interpretations is true or whether the agreement of the validity of interpretations is a sound one. It also makes it possible to wonder whether a useful interpretation is a true one. (See Lincoln & Denzin 1994, 580.) The importance of the correspondence theory of truth is tied to the aim of interpretation. If we want to find out the real reasons for action (e.g. the reasons that causally contribute to action), we are presupposing the correspondence theory of truth. The real reasons make reference to the mental processes involved in the control of action. The correspondence theory of truth does not necessarily state anything about the possibility of the actual comparison of interpretations and reality. In this sense it is possible to state the importance of the correspondence theory of truth for the understanding of the concept of truth or of truthful interpretation. Despite this one may be forced to depend, for example, on the coherence theory of truth in the actual evaluation of interpretations. The coherence, correspondence and consensus theories of truth do not necessarily conflict with each other. The correspondence theory states what the truth of interpretation means, and the coherence theory of truth states how the truthfulness of interpretations can be evaluated. The consensus theory of truth stresses the importance of the research community in the actual verification of interpretations. When we ask what the real reasons for action are, we must assume the correspondence theory of truth. Nevertheless, this question is not the only possible one. If we ask how far the agent s reasons for his action are rational, we are forced into a dialogue that concerns our own ideals and presuppositions, too. This dialogue presupposes the coherence theory of truth because we have to maintain a coherent world-view to be able to defend our own position. If the aim of research is to make reflection and emancipation possible in the community studied, we are trying to help people to find a better way of living. Here an instrumental theory of truth is supposed. Those interpretations that

13 Interpretation, Truth and Correspondence 389 make a deepened self-understanding or democracy in community possible are useful for the development of the community. If our aim is just to listen to the stories people are telling about their action and to understand these stories, we are engaged in a dialogue trying to build up a coherent understanding. The correspondence theory of truth forces us, for example, to wonder what the real reasons for action are. This is so because the reasons for action do not make a uniform case. In some cases we seem to lose touch with the reasons. It is not always easy to say what one has to compare the interpretations with. In fullblooded conscious action there is no problem, but the case of implicit reasons is not so clear. What is the ontological character of implicit reasons for action? Are there real implicit reasons that are not mental states? If so, what are they? Are they to be conceived of as logical constructs based on a description of action? Are they a part of background capacities? To my mind the conclusion is not that there is some vagueness in the notion of implicit reasons for action and furthermore that the correspondence theory of truth does not suit all cases of the interpretation of the reasons for action. The conclusion is, rather, that the correspondence theory of truth is important because it forces us to wonder what the ontology of the reasons for action is. Besides, as I have stated, the correspondence theory of truth is presupposed in many daily activities of qualitative research. Pentti Moilanen University of Jyväskylä Department of Teacher Education PO Box Jyväskylä FINLAND mope@edu.jyu.fi REFERENCES ALTHEIDE, D.L. & JOHNSON, J.M. (1994). Criteria for assessing interpretive validity in qualitative research. In: N.K. Denzin & Y.S. Lincoln (Eds.) Handbook of qualitative research (pp ). Thousand Oaks: Sage. ASHWORTH, P. (1993). Participant agreement in the justification of qualitative findings. Journal of Phenomenological Psychology 24, ATKINSON, B., HEATH, A. & CHENAIL, R. (1991). Qualitative research and the legitimization of knowledge. Journal of Marital and Family Therapy 17(2), BHASKAR, R. (1979). The possibility of naturalism. A philosophical critique of contemporary human sciences. Brighton: Harvester Press. CARLGREN, I. & LINDLAD, S. (1991). On teacher s practical reasoning and professional knowledge: considering conceptions of context in teacher s thinking. Teaching & Teacher Education 7, GERGEN, K. (1985). The social constructionist movement in modern psychology. American Psychologist 40,

14 390 Pentti Moilanen GRUSCHKA, A. & GEISSLER, H. (1982). Über die Fähigkeit von Untersuchten und Wissenschaftlern, interpretative Urteile zu validieren (The capability of researchers and persons studied to validate interpretations). Zeitschrift für Pädagogik 28, HABERMAS, J. (1973). Wahrheitstheorien (Truth theories). In: H. Fahrenbach & (Ed.) Wirklichkeit und Reflexion (pp ). Pfullingen: Günther Neske. HAMMERSLEY, M. (1992). What s wrong with ethnography? London: Routledge. HOWE, K. & EISENHARDT, M. (1990). Standards for qualitative (and quantitative) research: a prologomenon. Educational Researcher 19, 2 9. KOSKI, L. (1996). The truth, the quality, and the interpretation. In: K. Julkunen (Ed.) Qualitative methodology in educational research (pp ). University of Joensuu. Bulletins of the Faculty of Education. N:o 60. LENK, H. (1991). Zu einem methodologischen Interpretationskonstruktionismus (Towards a methodological interpretation-constructionism). Journal for General Philosophy of Science 22, LINCOLN, Y. & DENZIN, N. (1994). The fifth moment. In: N.K. Denzin & Y.S. Lincoln (Eds.) Handbook of qualitative research (pp ). Thousand Oaks: Sage. LINCOLN, Y.S. & GUBA, E.G. (1985). Naturalistic inquiry. Newbury Park: Sage. LONGINO, H. (1979). Evidence and hypothesis: an analysis of evidential relations. Philosophy of Science 46, MARTIN, R. (1976). The problem of tie in von Wright s schema of practical inference: a wittgensteinian solution. Acta Philosophica Fennica 28, MILES, M. & HUBERMAN, A. (1994). Qualitative data analysis. Second Edition. Thousand Oaks: Sage. NESPOR, J. (1987). The role of beliefs in the practice of teaching. Journal of Curriculum Studies 19, NOEL, J. (1993). Intentionality in research of teaching. Educational Theory 43, SANDBERG, J. (1997). Are phenomenographic results reliable? Higher Education Research & Development 16(2), SEARLE, J. (1992). The rediscovery of mind. London: MIT Press. SEARLE, J. (1995). The construction of social reality. New York: Free Press. SILVERMAN, D. (1994). Interpreting qualitative data. Methods for analysing talk, text and interaction. London: Sage. STRAUSS, A. & CORBIN, J. (1990). Basics for qualitative research. Grounded theory procedures and techniques. Newbury Park: Sage. TAYLOR, C. (1971). Interpretation and the sciences of man. The Review of Metaphysics 25, TERHARDT, E. (1985). The adventures of interpretation. Approaches to validity. Curriculum Inquiry 15, VOLLMER, F. (1995). The control of actions by agents. Journal for the Theory of Social Behaviour 25, WACHTERHAUSER, B. (1994). Introduction: Is there truth after interpretation? In: B. Wachterhauser (Ed.) Hermeneutics and truth (pp. 1 24). Evanston, Ill: Northwestern University Press. WHITMORE, E. & RAY, M.L. (1989). Qualitative evaluation audits. Continuation of the discussions. Evaluation Review 13(1),