Toward a Rational Reconstruction of Design Inferences
"From long habit the train of thoughts ran so swiftly through my mind that I arrived at the conclusion without being conscious of intermediate steps. There were such steps, however."
-- Sir Arthur Conan Doyle, A Study in Scarlet
Detecting the deliberate operation of intelligent agents in the environment is an activity so natural and pervasive that we may not be inclined, pre-reflectively, to classify it as an inference at all. The analogy with visual perception is initially very attractive. Just as we spontaneously discriminate among physical objects without any overt sense of deliberation, so we find ourselves sorting out signals from noise, extracting artifacts from archaeological digs, and mentally separating mechanical contrivances from their environmental contexts without weighty deliberation. It is a poor handyman who needs the help of philosophy to distinguish the lawnmower from the lawn.
But the detection of design also arises in contexts where, for a variety of reasons, the clarity of our vision shades off toward opacity. No doubt Spenser wrote the Amoretti: but did he deliberately space some of the sonnets so as to reflect the liturgical year, or is their spacing a coincidence? Does the feathered tail really go with the dinosaur body in this fossil, or are we being subjected to a clever hoax? Shall we classify the untimely demise of Birdy Edwards as a homicide or as an accidental death? No reasonable person doubts that in such cases we do well to pay close attention to a wealth of details that might not be apparent to the casual observer. Manuscript comparisons, radioactive dating, the patient sifting of forensic details -- the strategies we employ to disentangle design from chance are legion, and they make little sense if the detection of design is essentially a simple and unreflective matter of perception.
In the clear-cut cases, we often have little or no difficulty achieving consensus regarding design or chance: our common intuitions are robust. And even in the most complex cases where we are at a loss to explain the phenomena at hand, we are not wholly without rational guidance. Often we are able to see quite clearly what sort of evidence would, if only we could acquire it, strengthen or weaken the case for design. We have strong and stable intuitions regarding which aspects of our background knowledge are crucial and what sorts of new information would be relevant or irrelevant.
In view of the sensitivity of our judgments to available evidence, it makes sense to try to reconstruct our detection of the deliberate activities of intelligent agents as an inference -- to analyze both our pre-reflective practices and our more painstaking investigations with a view to making explicit their underlying rational structure. But before we embark on such a rational reconstruction, we need both a clear sense of what the project is about and some criteria by which to evaluate rival proposals. In the first section of this paper I discuss the potential tension between the rationality and the verisimilitude of our reconstructions and suggest that despite the difficulties inherent in this project the game is very much worth the candle. In the second section I propose seven criteria and argue that they are plausible constraints on any rational reconstruction of design inferences. In the next three sections I examine some current proposals regarding the methodology of design detection in light of these criteria, and in the final section I draw some conclusions from this examination.
[The remainder of this essay is omitted as it is under consideration at a journal. For a copy, please email me at mcgrew-at-wmich-dot-edu]