One of the things people do with texts is read them. This is certainly not the only thing people do with texts, nor, I would argue, is it the primary thing people do with texts, but it is one of the more fundamental things, and even when texts are used for other purposes-such as when one saves a cash register receipt in case one might need to return a purchased item-these other actions are carried out with an eye to the possibility, at least, of someone reading the text.

Scholarly approaches to the reading and interpretation of texts tend to fall into one of two broad tendencies. The first is the classic model of literary interpretation wherein the reader uses clues from the text to form ideas about what the text says. This model of reading tends to envision the single reader in isolation with the text and to focus on the features of the text that guide-well, should guide-the reader to the author's intended interpretation. This model tends to be highly normative, and I do not think was ever really intended to be a description of how people actually read.

The reaction to the classical model was a whole variety of reader-response theories that emphasize the active role of the reader in creating meaning. In some of these theories the text is almost entirely incidental. The anthropological versions of these theories have tended to emphasize discourse around texts, and to see reading as just a variety of social interaction. In these theories the text does nothing more than provide an occasion for interpretation, and the structure of the text is seldom discussed at all.

I've never been comfortable with either approach. I think that much about the reader-response approach is correct, provided that we understand reader-response theory as a psychosocial theory, not a literary theory: it is not a theory of meaning, but an outline of the formation of mental representations and their attributions, at least those involving texts. Anthropological studies examining the discourse around texts can be very helpful. James Bielo's recent Words upon the Word (Bielo, 2009) is an excellent example of an anthropological study that examines the formation of interpretations in small group Bible studies, where the formation of interpretations is partly collaborative and tied to the communities' self-definition.

But there is still something to be said for the classic, formalist model. People do read by themselves. People do use the text to reconstruct an intended meaning. People do not merely read into texts what they have been told it says. The structure of the text does matter for what people think about it. If one observes instances of interpretive discourse, one commonly-though not always-sees reference to the structure of the text.

There are three major problems with historical and literary studies of interpretation, as I see them:

• They assume that every interpretation is the result of interpretive activity. This is not so (Malley, 2004). Some ideas are attributed to a interpretand without ever having been derived from the interpretand. It requires historical investigation to determine how an interpretation was actually formed.

• They assume that the presence of a hermeneutic tradition, a normative tradition about the hermeneutic process that should be applied to the interpretation of a text is a description of what people actually do when they read the text. This is usually a tacit assumption, but it dominates studies of the history of interpretation. It is patently false, as most readers of texts are unschooled in the hermeneutic tradition and pay only selective attention to it when they do know it (Malley, 2004).

• They assume-and thus take for granted-the fact that we cannot predict interpretations on the basis of knowledge of the interpretand. We ought to find this a remarkable fact.

Any empirical approach to the study of interpretation must begin by discarding these assumptions.
We might begin to frame the problem of interpretation by asking whether it is in fact possible to predict interpretations. It seems like there must be some process such that, given the text, we can predict how people will understand it. After all, people do produce interpretations of texts without any input from sources other than the text, and as I write this I feel reasonably confident that I can anticipate how you, the reader, will understand what I am saying. If comments are posted, I will doubtless discover that some of this confidence is misplaced, but not that all of it is.

Stripped to its bare essentials, the problem of interpretive activity is just this: What, given a text, do people do to arrive at an interpretation? We can even express this symbolically, thus:

f (text) → interpretation

Our problem is to figure out f.

The basic conclusions I came to in my ethnography of evangelical Christian Biblicism were the following:

• f is a search process rather than an algorithm applied to the text. This conclusion has fared well in the face of more recent work, which shows that many conversations around the text amount to a collaborative search for an interpretive consensus.
• f has as its goal the establishment of a highly relevant connection. It is not true that Christian readers find in the Bible only what they want it to say, or use it only to reinforce pre-existing beliefs. Their reading is shaped by the expectation that the Biblical text will be highly relevant to them as individuals, but not by any particular notion of what that highly relevant connection will be.
• f permits the relation f(a) → -a. I observed an interpretive event where a small discussion group started with the text "All things are possible for God" and concluded that "not all things are possible for God." Although the explicitness of this contradiction is striking, the interpretive process that led to it was not different than the interpretive processes that led to other, less remarkable conclusions. I believe that this sort of contradiction in fact happens very often, but is seldom explicitly articulated. In any case, the fact that f permits this relation means that f is probably not itself an explanation of the set of interpretations present in a community, even those that were formed out of interpretive processes.

I hasten to add that these conclusions were formed from a sample of data drawn from a single interpretive community. There is no reason to assume that f is identical from one individual to another, much less from one community to another, and still less from one tradition to another. I regard f as almost completely unspecified despite my best efforts, though I believe that I have circumscribed it in an interesting and productive way.

Whether the interpretive process in other traditions may be similarly circumscribed is an open question. If I may speculate freely, I think we will find, when traditions are examined systematically and empirically, that these conclusions are true of nearly all interpretive traditions. It is possible for these things not to be true of an interpretive tradition, but I believe that any such tradition would incur enormous costs in terms of energy and the restriction of interpretations, and would end up having to be quite elitist in order to maintain itself. The popular tradition, I conjecture, would revert to something analogous to the Biblicist tradition.

What about non-textual interpretands? With non-textual interpretands, there is no possibility of finding f(a) → -a, just because a is not propositional in nature. The other conclusions, however, could still be true of a non-textual interpretand. To return to the example of a martial arts form, typically interpretations take the form of ascribed health benefits or combat applications of the forms. Many of these are more a process of attribution than interpretation, so we will look at them more closely next month, when we consider the role of attributions in interpretive traditions.

Bibliography

Bielo, James S. (2009). Words upon the Word: An ethnography of evangelical group Bible study. New York: New York University Press.

Malley, Brian E. (2004). How the Bible works: An anthropological study of Evangelical Biblicism. Walnut Creek, CA: AltaMira Press.