What is it we intend that a theory should do for us? Is it meant to be a distillation of basic truth? Is it meant to be a fruitful guide to research. Is it meant to be a specific statement about the way things work?

It might be useful to distinguish between these different tasks, or between different kinds of theoretical formulation that serve them. One much discussed distinction is between highly abstract, general theory, and more specific, lower level theory. A famous sociologist of the mid-twentieth century, Robert K. Merton, favoured the latter, labelling it “middle range theory,” in contrast with the more abstract “grand theory.” How did he define, and why did he favour “middle range theory”? Middle range theory is specific enough to say what can or can not happen, and therefore can be tested by evidence. Grand theory is so abstract as to make it difficult to challenge with evidence.

Most “theory” in anthropology is highly abstract. I call it “heuristic theory,” because it is mainly meant to guide research by indicating what is important. Heuristic theory is more or less equivalent to grand theory, although sociologists, e.g Talcott Parsons and C. Wright Mills, tended to elaborate more than anthropologists do. Perhaps a paradigmatic example would be Clifford Geertz’s theoretical statements about religion, common sense, art, etc., which are frank in taking the form of definitions. Heuristic theory, and definitions, cannot be shown to be true or false; rather they are seen as either fruitful or not (or, "old" or "new," as fashion is our master, even in anthropology).

In contrast to heuristic theory are formulations that I would label “substantive theory,” which are more or less the same as Merton’s middle range theory. These tend to be fairly specific about something in particular, such that evidence can be brought to bear in support or in refutation. A simplistic example in one proposition would be “nomads do not have kings.” This substantive theory could be addressed by looking at as many cases of nomads as we can find, and seeing if any have kings. (My bet: don’t put your money on the kings.)

In addition to heuristic theory and substantive theory, there is epistemological theory, that asserts how we can know things, and may support or dismiss heuristic theory or substantive theory, or both. Usually epistemology is attended to by philosophers, but as nothing human is alien to anthropologists, we dabble in epistemology as well.

Views: 2084

Replies to This Discussion

Should a theory be applied to fit the data or the fieldwork driven to fit a theory?
More importantly, what are we using theory for, of what do we want to use it for?

Experience and opinions will vary on this point, of that i am sure!
Regarding your first question, I think both go hand in hand. Fieldwork is driven, in part, by the theoretical disposition of the anthropologists, so in a way, fieldwork is always driven to fit a theory (or more likely, an amalgam of theories). However, the anthropologists isn't a passive fieldworker, and the encounters they make may result in a change to their theoretical posture. Theory always filters the data we find, but I think data can also reveal the limited scope of any single theory, requiring anthropologists to keep a flexible theoretical toolkit.

Could you provide an example or two in which, "Epistemological theory is especially important to anthropology, as it stands right now, because it discusses knowledge"? Methodological discussion with a strong operational, i.e., how we might do this, component—that I can see the value of. But "epistemological theory"? I'd like to see a demonstration.
I have a hard time with social constructionism when it comes to questions of other knowledge systems, for example the Lakota Sioux and the vision quest. Are they really receiving sacred knowledge through the vision quest or should I try and explain the ritual through other means? I'm inclined not to discount what they believe, partially out of respect for their beliefs but also partially because I just can't know.

Do the answers to the anthropologist's questions depend on whether the anthropologist shares the views of those whose lives we share and study? Here is another possibility, from a paper I published in american ethnologistVol. 22, No. 1, February 1995.

I start with a working assumption, nicely stated by Erving Goffman, that applies as well to the presentation of self in ritual as it does to everyday life.
When an individual plays a part he implicity requests his observers to take seriously the impression that is fostered before them. They are asked to believe that the character they see actually possesses the attributes he appears to possess, that the task he performs will have the consequences that are implicitly claimed for it, and that, in general, matters are what they appear to be.

I assume, then, that the healer is doing, in fact, what he seems to be doing: negotiating with demons. He is neither charlatan, preacher, nor pedagogue; nor is he an actor performing a play that he and others know to be fiction. He is what he says he is: a magician, trying to achieve a certain effect in the way he knows best, by magic. This leads me to my larger agenda, the implications of what he says for one of our discipline's oldest conundrums: the magical force or efficacy attributed to magical words.

The paper then goes on to examine the language used in this particular example of a common popular Daoist exorcism in light of three anthropological theories:

1. That the ritual is, in some sense, what J.L. Austin called a performative act.
2. That the ritual is, as James Fernandez suggests, the embodiment of metaphors that shove around the various actors in the rite in a multidimensional cultural space.
3. That, as Maurice Bloch argued, magical language is formalized language, deliberately making use of syntactic and lexical restrictions to assert unassailable authority.

The relevant conclusions are

1'. Most of the ritual is, in fact, a prologue to a final performative act: a process of persuading the amoral demons to accept the social contract that empowers the magician to exorcise them.
2.' That Fernandez' approach complements the idea that rituals are social dramas; the metaphors in play first position the actors then move them to a desired place. The magician, for example, starts out as a neutral negotiator who treats the demons with a moderate amount of respect. Then, as the climax is reached, he identifies himself with the gods and reduces the demons to powerless anonymity.
3.' That the degree of formality with which the magician speaks varies as the ritual proceeds. Bloch is right in that the language is most formal when authority is being most strongly asserted; but this is not always the case. When, for example, the magician is tossing the divining blocks and tossing more spirit money into the fire as he looks for the demons' agreement to the deal he is offering, his tone is highly informal.

These results point to the major conclusion that the three theories in question are all useful in raising questions about what is going on. They are not, however, mutually contradictory; nor is one plainly superior to the others. They are, as it were, like the various perceptions of the blind men touching the elephant. They all tell us something, not everything.

The point I would add here is that none of this analysis requires taking a position on the reality or lack thereof of the demons and gods invoked in the rite. It suffices to take the magician seriously and examine respectfully what he says -- to treat him in other words as we would treat any colleague.
Thanks for this enlightening contribution, John. I would like to endorse your main conclusion: theories are "all useful in raising questions about what is going on. They are not, however, mutually contradictory; nor is one plainly superior to the others...They all tell us something, not everything."

I tell my students that any theory that has been around for a while -- has achieved a measure of social acceptance, however narrow -- must be good for something. That something is unlikely to be as wide-ranging or universal as its protagonists claim for it and its value must be seen as being relative to the questions being asked and to complementary ways of answering them.

When I was an undergrad, I was a Durkheimian, as a PhD student a Weberian, then as a young professor and for the longest time a Marxist. I found Hegel in the Caribbean and Kant soon thereafter. By the time that my favourite thinker was Rousseau I began to figure out where I was heading and, true enough, when I wrote my book on money, the most cited author was John Locke, progenitor of the liberal Enlightenment. I stopped there. Thomas Hobbes, pace Marshall Sahlins, was a step too far. I have now lived in Paris for 12 years and next week will be co-organizing a big international conference on Marcel Mauss, who is my present hero. But what strikes me about that intellectual journey to Locke and back again is that I found it necessary to focus on one of them at a time. Now I can raid the whole period eclectically and truly live up to my pedagogical formula. I am me and I can draw on whoever or whatever I like, according to the question and context.
When we can say, as John McCreery does, that
“the three theories in question are all useful in raising questions about what is going on. They are not, however, mutually contradictory; nor is one plainly superior to the others.”
we are clearly dealing with heuristic theories. Heuristic theories point out things that they claim are important. So different heuristic theories that point out different aspects of human life can be complementary and equally useful.

It is obvious that you could not say the same about substantive theories specifying that something is the case and something is not the case. The following three substantive theories are mutually contradictory, and cannot all be equally useful. In fact, we do not ask whether they are useful; we ask which one is correct:

Theory I: Nomads do not have kings.
Theory II: All nomads have kings.
Theory III: Nomads in tribes larger than 10,000 have kings.

Our decision about which one is correct, or closest to correct, is based on the information we have about nomads. By information I mean data or facts that serve as evidence that nomads are one way or another. In this way, substantive theories can be tested and supported or refuted, or, in extremity, modified. At the end of the process, we have a proposition that states what nomads are like and not like, and that has so far been supported by the evidence. Contradictory propositions has been shown to be inconsistent with the evidence, and thus incorrect.
Two points: (1) While thanking Keith Hart for his kind words, I don't feel comfortable generalizing what I said about three particular theories in relation to one particular case to all theories in all cases. In part this is (2) because I largely agree with Philip's distinction between heuristic and substantive theories. That said, I would be inclined to agree even more if, given the examples in question, we were simply drawing a distinction between heuristics (which suggest moves toward understanding) and hypotheses (testable substantive claims). I would tend to see Phil's three "theories" (which I would call hypotheses) as possible candidates for inclusion in a theory conceived as a body of propositions deducible from principles supported by a variety of evidence provided by either (a) experiments, (b) statistical analysis, or (c) qualitative research that satisfies basic requirements of including alternative hypotheses and evidence that supports one more strongly than the other. If forced to decide between two hypotheses based on evidence provided by qualitative research, I would favor the one that accounts for more detail and the sequence of events that the evidence in question describes.

It strikes me that "theory" is, like similar terms, e.g. "values," "law," "economics," "politics," "art," or "philosophy" a polyvocal symbol, with a range of possible and sometimes contradictory meanings. We could choose one by fiat (as I have done in the paragraph above) or adopt what is to me a more interesting and anthropological approach by mapping those possible meanings and exploring their relation to the contexts in which they occur.
John McCreery: I would tend to see Phil's three "theories" (which I would call hypotheses) as possible candidates for inclusion in a theory conceived as a body of propositions deducible from principles supported by a variety of evidence provided by either (a) experiments, (b) statistical analysis, or (c) qualitative research that satisfies basic requirements of including alternative hypotheses and evidence that supports one more strongly than the other.

Of course, John, I entirely agree. The “substantive” example was oversimplifed. The propositions, or hypotheses, offered would have to be part of a general theory about leadership and its bases.
I am glad of the clarification reached through the last two posts. But I still think that there is a wide range of levels of generality that could be attached to the word "theory" and that could lead to misunderstanding between us, probably already has on my part.
Let me give an example of what I think of as theory and its eclectic application. I am not impressed by the heuristic/substantive pair, but perhaps the two of you can show me how it applies to the following.

I have an undergraduate lecture on capitalism, which is itself an ideal type of some generality, even theoretical significance. Marx defines it as "the exchange of free capital for free labour" and in Part 8 of Capital Vol 1 uses that definition to provide a history of "primitive accumulation" or where capitalism came from. Weber defines it as "rational enterprise" and in General Economic History Part 4 explains the rise of capitalism as developing the social and cultural conditions for calculating uncertain futures.

I then take David Parkin's Palms, Wine and Witnesses, an account of incipient capitalism in Coastal Kenya, and show how the ethnography can be understood partly through the application of Marx's scheme (development of a wage labour force and of property rights in coconut trees), partly by using Weber (the entrepreneurs escape from the ruinous expense of a drinking culture by becoming Muslim).

Now a Marxist or a Weberian would claim that Giriama proto-capitalism could best be explained by their focus alone; and this is what I meant by the claim that all theories have more partial relevance than their adherents usually claim. [I could use the same argument for a dispute over say Leach's, Turner's and Bloch's theory of ritual].

We are not talking here about testing a scientific hypothesis, but about competing stories concerning what capitalism really is and implicitly what to do about it if you like or don't like it. Have I been talking about theory in your terms at all? How does the heuristic/substantive pair apply?
You are right, Keith. Heuristic theories are stories that emphasize particular aspects of life and direct researchers to pursue these aspects in their research. R-B said life is about ongoing social relations: Look at them. Geertz said life is all about meaning: Look for meaning. Barth said life is about choices, decisions, and transactions: Look at them. These and others go in and out of favour, but we could certainly draw on more than one. What we cannot do is say that our studies have shown that one or the other of these heuristic injunctions is false.

As you know, Karl Popper argued that scientific theories have a particular characteristic: they specify that certain things must happen and others not. If what they say happens, the theory is supported. If what they say should not happen, happens, then they are refuted. Robert K. Merton intended “middle range theories,” or, as I have called them, “substantive theories,” to be of this type.

An example of a substantive theory is Robert L. Carneiro’s argument that indigenous technological and political development took place in “regions where the area of cultivable land was distinctly circumscribed.” These developments, according to Carneiro, did not take place in broad, open regions. To assess this theory, we must survey cases of indigenous societal development, so see if the predicted pattern is sustained by the evidence. If we find as many, or even many cases of development in open regions, then Carneiro’s theory would be refuted.

Popper knew that there were many “stories” about the social world, including those of Marx and Freud, to which he gave considerable attention. But because any fact that could be found, could be explained one way or another in a Marxist or Freudian fashion, these could not be considered scientific theories. In the psychoanalytic tradition, a person’s response to an event may be negative or positive, but can somehow be explained or at least accommodated by the theory. Popper rejected the value of such all encompassing stories. For him, no real knowledge, understanding, or explanation could come from them.

For myself, I don’t mind using unscientific heuristic “theories” as stimulants for research and analysis. But I do think that we are obliged to go beyond them to substantive theories that are risky formulations subject to the evidence of our ethnographic discoveries.
Is the difficulty partly that the ideas we describe as theories can be either heuristics, hypotheses or, very commonly, both. Marx, for example, can be read as offering a substantive theory one of whose major hypotheses is the inevitable failure of capitalism when the contradiction between capitalists and workers leads to the revolution. Marx can also be read as heuristics, e.g., (1) start with the definition and distribution of property, (2) note that different interests will, if unchecked, inevitably lead to conflict, (3) map the current state of play, trace as much of the history as you can, identify the principal actors. You may, then, (4) achieve a deeper understanding of how capitalism or, indeed, any social system works. The heuristics may still be valuable even if the hypothesis remains unconfirmed.

It occurs to me that hypothesis versus heuristic distinction is closely analogous to the 'models of' versus 'models for' distinction that Clifford Geertz introduced in his essay, "Religion as a Cultural System." In our, as yet totally imaginary, best of all possible worlds, the distinction would collapse; the two models would seem only different aspects of the same integral whole. The What would imply the Why, the fact the value, the hypothesis the heuristic, the testable claim the course of action. In our actual, imperfect world, gaps appear, creating spaces for scholarly and political debate.

The trick is not to fuss about whether a proposition is or is not "essentially" an hypothesis or heuristic, but instead to look pragmatically at how it is being used. If an author is offering evidence to test its validity, call it an hypothesis. If the author is extracting from it interesting questions to pursue, call it a heuristic. The difference lies not in the raw proposition itself, but rather in what we do with it.

WARNING: Like all arguments conceived in terms of binary oppositions, this one is probably simplistic. I look forward to suggestions on how it might be further improved.



OAC Press



© 2019   Created by Keith Hart.   Powered by

Badges  |  Report an Issue  |  Terms of Service