ACM Interactions

Theory weary, theory leery,
why can't I be theory cheery?
I often try out little bits
wheresoever they might fit.
(Affordances are very pliable,
though what they add is quite deniable.)
The sages call this bricolage,
the promiscuous prefer menage...
A savage, I, my mind's pragmatic
I'll keep what's good, discard dogmatic…

—Thomas Erickson, November 2000
"Theory Theory: A Designers View" (sixty-line poem)

An attentive reader of my blog posts on bias and reverse engineering might have noticed my skirmishes against the role of theory in human-computer interaction. I’m losing that war.

Appeals to theory are more common in some fields than others. CHI has them. CSCW has more and UIST has few. I’m writing on the plane back from CSCW 2014, where we saw many hypotheses confirmed and much theory supported. In one session I was even accused of committing theory myself, undermining my self-image of being data-driven and incapable of theorizing on the rare occasions that I might like to.

One author presented a paper informed by Homophily Theory. He reported that it might also, or instead, be informed by Social Identity Theory. After reading up on both, he couldn’t tell them apart. So he settled on Homophily Theory, which he explained meant “birds of a feather flock together.” It was on the slide.

When I was growing up expecting to become a theoretical physicist, “birds of a feather flock together” was not considered a theory; it was a proverb, like “opposites attract.” I collected proverbs that contradicted each other, enabling me to speak knowingly in any situation. Today, “opposites attract” could be called Heterophily Theory, or perhaps Social Identity-Crisis Theory.

In the CSCW 2014 proceedings are venerable entries, such as Actor-Network Theory and Activity Theory. The former was recharacterized as an “ontology” by a founder; the latter evolved and is considered an “approach” by some advocates, but we don’t get into them deeply enough for this to matter. Grounded Theory is popular. Grounded Theory covers a few methodologies, some of which enable a researcher to postpone claiming to have a theory for as long as possible, ideally forever. But some papers now include an “Implications for Theory” section; as with “Implications for Design” in days of old, some reviewers get grumpy when a paper doesn’t have such a section. With CSCW acceptance rates again down to around 25%, despite a revision cycle, authors can’t afford to have grumpy reviewers.

CSCW citations also include broad theories, such as Anthropological Theory, Communication Theory, Critical Theory, Fieldwork for Design Theory, Game Theory, Group Dynamics Theory, Organizational Science Theory, Personality Theory, Rhetorical Theory, Social Theory, Sociology of Education Theory, and Statistical Mechanics Theory. (These are all in the proceedings.) Theory of Craft is likely broad (I didn’t look into it), but Theory of the Avatar sounds specific (didn’t check it out either).

The Homophily/Social Identity team did not get into Common Identity Theory or Common Bond Theory, but other authors did. I could explain the differences, but I don’t have enough proverbs to characterize them succinctly. With enough time one could sort out Labor Theory of Value, Subjective Theory of Value, Induced Value Theory, and (Schwartz’s) Value Theory. All are in CSCW 2014, though not always explained in depth. So are Resource Exchange Theory, Social Exchange Theory, Socialization Theory, Group Socialization Theory, Theory of Normative Social Behavior, and Focus Theory of Normative Conduct.

We also find models—Norm Activation Model, Urban Gravity Model (don’t ask), Model of Personal Computer Utilization, and Technology Acceptance Model. The latter has a convenient acronym, TAM, giving it an advantage over the related Adoption Theory, Diffusion of Innovations Theory, and Model of Personal Computer Utilization: an Adoption Theory acronym would risk confusion with Activity Theory and Anthropological Theory, and who wants to be called DIT or MPCU? Actor-Network Theory has a pretty cool acronym, as does Organizational Accident Theory—both acronyms are used.

Although they don’t have theory in their names, Distributed Cognition (DCog) and Situated Action are popular. Alonso Vera and Herb Simon described Situated Action as a “congeries of theoretical views.” Perhaps in our field anything with theory in its name isn’t really a theory.

Remix Theory and Deliberative Democratic Theory sound intriguing. They piqued my interest more than Communication Privacy Management Theory or Uses and Gratifications Theory. The latter two might encompass threads of my work, so perhaps I should be uneasy about overlooking them.

The beat goes on: Document Theory, Equity Theory, Theory of Planned Behavior (TPB). CSCW apparently never met a theory it didn’t cite. There is also citation of the enigmatically named CTheory journal. What does the C stand for? Culture? Code? Confusion?

Graduate students, if your committee insists that you find another theory out there to import and make your own, find an unclaimed proverb, give it an impressive name, and they’ll be happy. Practitioners, what are you waiting for, come to our conferences for clarity and enlightenment!

***

Postscript: This good-natured tease has a subtext. Researchers who start with hypotheses drawn from authoritative-sounding “theory” can be susceptible to confirmation bias or miss more interesting aspects of the phenomena they study. Researchers who find insightful patterns in solid descriptive observations may suffer when they are pressured to conform to an existing “theory” or invent a new one.

Thanks to Scott Klemmer for initiating this discussion, and to John King and Tom Erickson for comments.

Jonathan Grudin