Lately, a lot of my intellectual energy has been devoted to issues that one way or another pertain to the notion of the hermeneutic circle. The following is an effort to develop some clarity on it.
Defining the Hermeneutic Circle
Just so everyone knows what I am going on about here, let me first attempt a definition, and then I’ll develop the concept in two different directions. The hermeneutic circle refers to the situation in which when we encounter a text (i.e., any cultural phenomenon) we can only understand it (i.e., make sense of it) with reference to other texts, and in turn our understanding of these other texts is modified by our understanding of this text.
In other words, to understand a given expression, we must understand the language in which it is written. “Language” here means more than English or Chinese. For example, to understand an expression such as Shakespeare’s Antony and Cleopatra, one must have some competence with Western tragedy in general, Shakespearean tragedy in particular, Roman history, poetry (e.g., meter), Renaissance culture, and so forth. Yet these linguistic contexts are all made out of other texts, which can only be understood the same way. We face just such a problem in HCI when we interpret and/or evaluate an interaction: the criteria by which we evaluate or interpret an interaction are determined by concepts and theories ingrained in our field, and these in turn are derived from, among other things, interpretations and evaluations of other interactions.
In this post, I want to explore two takes on the hermeneutic circle, a pessimistic view and an optimistic view. I will also make at least token efforts to relate this discussion to HCI.
Take 1: The Pessimistic View
The hermeneutic circle seems to interfere with, if not absolutely preclude the possibility of, objective knowledge. This is so because our reasoning would seem to be circular. The only way to justify a reading (by which I mean “serious act of sensemaking”) is to point to other readings, which themselves need justification in the same way, leading to an infinite regress argument.
Another practical implication is that it is hard to enter one of these circles, since every possible entry point seems to take for granted the whole. You can’t understand Shakespeare without being in a position to understand Shakespeare; how can you get in such a position without reading Shakespeare?
Since the hermeneutic circle, on this view, is bad, we should try to break out of it. Philosopher Charles Taylor, in his essay “Interpretation and the Sciences of Man” explores this argument. He writes that one way people have tried to break out of it is by taking recourse to empirical science. Empiricism is, he writes,
a genuine attempt to go beyond the circle of our own interpretations, to get beyond subjectivity. The attempt is to reconstruct knowledge in such a way that there is no appeal to readings or judgments which cannot be checked further. That is why the basic building block of knowledge on this view is the impression, or sense datum, a unit of information which is not the deliverance of a judgment, which has by definition no element in it of reading or interpretation, which is a brute datum. The highest ambition would be to build our knowledge from such building blocks by judgments which could be anchored in a certainty beyond subjective intuition…. By “brute data,” I mean … data that cannot be questioned by offering another interpretation or reading, data whose credibility cannot be founded on or undermined by further reasoning.
On this view, empiricism (“let the data speak for itself”) offers an alternative to the hermeneutic circle, a way of knowing that is not entangled in it. We see an example of this sort of reasoning in Csikszentmihalyi & Robinson’s The Art of Seeing:
An analysis of any sort begins with a description of the phenomena under study. Yet a thoroughgoing and empirically grounded description of the aesthetic experience has been conspicuously absent from aesthetic theory–of whatever stripe–in the past. For the most part, aesthetic study has proceeded from a priori assumptions concerning what the aesthetic experience must be or the basis of the analyst’s own experiences.
So the authors are promising to do what 2,500 years of aesthetic theory have failed to do, namely, offer a rigorous description of the aesthetic experience (because apparently only empirical study is “thoroughgoing” and gets past “assumption”). They promise to replace empty theorizing with good, empirical evidence.
But pure empiricism has problems. As Wittgenstein, Quine, and other philosophers of science in the twentieth century have demonstrated, pure empiricism does not work, even in physics, let alone the human sciences. A hypothesis cannot exist without a theory already in place to give it coherence: How is the phenomenon studied? How is its data analyzed? On what basis is this hypothesis deemed important in the first place?
Additionally, Taylor notes, the brain simply doesn’t work that way to begin with, that is, pure empiricism itself entails a theory of perception that is outmoded. Finally, Taylor observes that for this kind of empiricism to work, reality itself must be structured in a certain way for it to be available to this kind of perception and knowledge building, i.e., it must be made available in a comprehensive and truthful way to human perception.
So how did Csikszentmihalyi & Robinson do? Using a grounded theory approach (note that grounded theory is based on a priori assumptions about what empirically informed theory must be), they interview nearly 70 museum curators (this “population” was “selected” based on established social science methodologies). So before they interviewed a single subject, their study itself is situated within the hermeneutic circle of empirical social science. How about their data? Their analysis of the data revealed that the curators express themselves according to four interrelated dimensions of aesthetic experience: perceptual, emotional, intellectual, and communicative. This may be surprising, inasmuch as it looks more like the table of contents of a psychology textbook than a work on aesthetics; the authors have minimally delivered on their promise of not revealing aesthetics as usual, but it seems fair to question whether they instead delivered psychology as usual.
How about the “brute data” itself? The authors offer quotes about how the curators perceived art. Here is an example: one curator describes the “color and forms,” “paint manupulation,” “dry, chalky lines,” and “stuccolike surfaces” of a painting; that sounds perceptual. The problem is interlaced among these terms were references to this work as “a perfect Cubist picture” and “that you associate with Cubist paintings.” In short, perception itself, by Csikszentmihalyi & Robinson’s own account was mediated by the hermeneutic circle: the painting was rendered perceptible to the curator on the basis of its participation in the Cubist tradition. Of course, the “Cubist tradition” refers to a bunch of subjective theorizing by artists, critics, and aesthetic philosophers; it certainly isn’t an empirically derived category. In short, Csikszentmihalyi & Robinson have not gotten out of the hermeneutic circle, as promised in their rhetoric; all they have done was empirically demonstrate that museum curators are themselves inside of a hermeneutic circle (i.e., art history), suggesting that all those dusty “a priori assumptions” that they were seeking to circumvent in fact are the basis of aesthetic experience.
Now, HCI is an empirical field, and its rhetoric sometimes resembles that of Csikszentmihalyi & Robinson. But as I have argued here, empiricism does not exempt anyone from the hermeneutic circle. Perhaps HCIers will resist this, but evidence of the presence of the hermeneutic circle in our field abounds. First, the idea that HCI is a “field”–what does that mean, if not that we all share a common general understanding (i.e., theory) of what HCI is? Such a shared theory is necessary for degrees in HCI to have meaning, for job descriptions to recruit the right sorts of people, and so on. We prefer certain kinds of research processes: empirical research is privileged over critical essays, for example. Equally, we as a field prefer certain research methods and epistemologies over others. Second, imagine a CHI paper with no references. With a pure model of empiricism, references would not only be superfluous, they would be corrupting! Yet how many of us when we submit are castigated for not having this-or-that paper in our references? This is an assertion of the hermeneutic circle (along with a second assertion that our work is outside of it). Or consider a criticism about an inappropriate reference: such a criticism amounts to a claim that the cited reference or definition doesn’t belong in the hermeneutic circle, or at least not in the way presented.
So it seems that the pessimistic view is well named; it views the hermeneutic circle in negative terms, and it is unable to get out of it.
Take 2: The Optimistic View
The optimistic view would seem to start with a few basic positions. First, pure empiricism is impossible and the hermeneutic circle is inevitable, and yet science has not faltered and dissolved. It seems that scientific progress is possible even if it is stuck within the hermeneutic circle. So, it may be that Csikszentmihalyi & Robinson’s claims to a privileged understanding of the aesthetic experience were delusional, but nonetheless they obviously offer genuine insight on how curators experience art, and in particular they reveal (even if inadvertently) how important aesthetic theory is to art appreciation, and in which sorts of ways. Above all, they help us see how a model audience appreciates art, a contribution to knowledge that benefits art educators, museum exhibit designers, art industry marketers, art critics, and so on.
A second benefit of a given field’s hermeneutic circle is that it renders itself visible. Let me explain that. Anyone remotely familiar with ubiquitous computing knows that Mark Weiser’s “The Computer in the 21st Century” was a seminal essay; everyone cites it and many people define themselves in relation to it. Because we know that this paper is a formative contributor to ubicomp, (a) we know that we need to teach this essay to graduate students in HCI; and (b) we know that anything obsolete, wrong, or brilliantly right about that essay are likely to have repercussions in ubicomp. In short, recognizing the hermeneutic circle(s) in which you operate de-naturalizes your work and renders it visible to reflection and critique. That in turn may prevent you from making dumb claims like you are going to fix the error made by 2,500 of years of philosophy going back to Plato by having your graduate students interview 67 people, and instead making claims about what your contribution actually is (e.g., help making visible how model or expert art viewers experience art).
The essence of the optimistic view, though, goes back to Gadamer (and was inadvertently given support in Csikszentmihalyi & Robinson). For Gadamer, our participation in the hermeneutic circle extends our perceptual horizons. Not only can I perceive what my senses make visible, but I can perceive the world making use of concepts that I have acquired by participating in a hermeneutic circle. We see this in the Csikszentmihalyi & Robinson paper where a curator expresses what she or he perceived in terms of Cubism. Without knowledge of Cubism, this curator may not have even noticed the chalkiness or stucco-texture of the paint. The perceptual qualities of the paint are only visible, significant, and interesting when viewed with an understanding of Cubism.
This line of reasoning obviously applies to HCI. I don’t have to reinvent “usability”; I can read about it, understand what it is, why it matters, how it has been evaluated, and it becomes a part of the repertoire with which I view, interpret, and experience a design. The concept of usability, therefore, is like an extra set of eyes, a new sense. Using Dewey’s and later anthropological notions of “experience,” I am able to reflect more thoughtfully, precisely, and substantially on my own experiences with interaction designs. This also means that HCI does not have to reinvent aesthetics, but rather can avail itself of aesthetics’ existing tradition, using it to extend our capacity in HCI to engage with beautiful, artistic interaction.
In short, the hermeneutic circle makes it possible for me to see and to participate in HCI as a field. It enables me to leverage the contributions of thousands of people in my own work, amplifying–or rather creating the very possibility of–my capacity to produce knowledge in the field. It also offers my colleagues a basis on which to critique and thereby help me improve my work. Finally, it offers all of us the opportunity to critique not just seminal papers, but the rippling of their effects across the field. In the philosophy of science, the realization that objective empiricism does not work (at least in a pure sense) led to the development of theorizing around scientific practices (such as triangulation) that seek to address the intrinsic weaknesses of pure empiricism in a way that allows empirical approaches to move forward on surer footing.
The hermeneutic circle is not a trap to avoid, but rather an opportunity to participate in the constructive development of our field. Our field, like all scientific fields, is hermeneutic, but that does not mean it cannot continue to rely on empirical strategies. It merely means that we should not be delusional about how this knowledge is produced.