Source
- Kendler (2008), LIT-539, read in full in NOTE-432 from the NIH author manuscript. Its section on “The Nesting of Biological and Psychological Explanatory Perspectives” is the whole argument for this account.
Kendler 2005 (LIT-552) is the paper the 2008 one cites for its premise that etiological models must be pluralistic. Its abstract asserts mind–brain causation in both directions and argues against dualism and epiphenomenalism. It is filed unread, so nothing here rests on it.
What was actually shown
Kendler (LIT-539) distinguishes two relations a lower-level theory can bear to a higher-level one (NOTE-432, C6). In replacement the higher-level construct becomes redundant, as temperature is once it is mean kinetic energy. In implementation the lower-level theory supplies “the mechanistic details of how the functions proposed by the more abstract theory actually get accomplished”. Marr’s levels give the order: function is stated first, and it can only be stated in psychological terms (“Neurochemical terms will not work”). A gene-to-protein finding has not explained schizophrenia until someone can say how the abnormal protein produces delusions and hallucinations.
The evidence offered is historical. Following Gold and Stoljar on Kandel’s Aplysia work, Hatfield on vision and Bechtel on memory, neuroscience has advanced by implementing processes psychologists described first. Kendler concludes that biology and psychology “will coevolve”. He also concedes that hard-reductive and hard-emergent work can still be useful: therapies have come from gene findings without a mechanism, and CBT “emerged from psychological constructs that contained no biology”.
What could have come out otherwise is the historical record: the cited cases could have been ones in which the psychological construct was dropped. None of the cited cases is psychiatric, and the record has read none of them. That is the reason for Proposed. Kendler himself grants that Marr’s scheme was built for information-processing systems and that “many psychiatric problems are not currently amenable” to that perspective, then holds that its logic is “valid for psychiatry across the board”. That step is asserted.
How it bears on mental causation, levels and emergence
- It is a thesis about explanation, not about extra causes. Pernu (LIT-147, NOTE-091) argues that the neural evidence for mental causation supports it only by identifying the mental with the neural: what is multiply realised in brain–computer interfaces is a coarse neural variable, not the intention. Kendler’s implementation relation does not need the psychological level to be a distinct cause, or to be multiply realised. It needs only that the function be stated psychologically and that the biology be how it is done. That fits Pernu’s deflationary option, in which mental causes are physical causes, as well as a nonreductive one. What it denies is that the psychological description drops out of the explanation, which Pernu’s identity reading does not claim either: “physicalism is not microphysicalism”. NOTE-432 makes the same point, that implementation is a non-eliminative alternative to the framing Pernu argues within. So the two sources are compatible, and the record holds no argument that makes psychiatric explanation choose between them.
- Levels are part–whole, and that limits “top-down” talk. Kendler treats levels as part–whole relations within one mechanism, and lists “recursive, or top-down, causality” among the features that defeat decomposition. Hazelwood (LIT-156, NOTE-137) argues from Craver and Bechtel and from Woodward’s independent-fixability condition that a whole and its own parts cannot stand in a causal relation. Kendler’s examples pass that test, because they run through time and through other people: tolerance raising later drinking, offspring of heavy drinkers abstaining. They are diachronic loops between distinct variables, not a whole acting on its parts at an instant. That reading is the record’s, not Kendler’s.
- It is neither hard reduction nor hard emergence. Kendler rejects both by name (NOTE-432, C4). The emergence the record holds from physics (Rizi, LIT-150; McKenzie, LIT-141) is weak emergence by coarse-graining, which is close to what Kendler’s local decomposition and reassembly assumes. Neither physics paper discusses minds.
- It is functionalism about psychiatric explanation. Stating a function first and its realiser second is the functionalist premise that THEORY-043 runs in two directions for minds and groups. Kendler applies it to explanation only, and says nothing about consciousness.
What this does not say
- It does not say biology is unimportant. Implementation is a demand that biological findings be connected to function, not a claim that they are secondary.
- It does not say the psychological level causes the biological one. Kendler’s “top-down” loops are cross-level in his sense, but they need no causation by a whole of its own parts (above).
- It does not cover the 2005 paper’s claims on first-person experience, dualism and epiphenomenalism. LIT-539 does not argue them.
- It does not settle what a psychological function is for psychiatry. Functional contextualism, the philosophy behind acceptance and commitment therapy (LIT-555, NOTE-431), limits causal analysis to manipulable context and says thoughts do not cause actions “except as regulated by context”. On that view the explanatory level is a person–context relation, not an internal function, so it would not accept Kendler’s Marr-style ordering as stated. The two have not been compared in print in anything the record holds.
Connections
- THEORY-058, on self-reinforcing loops, takes Kendler’s cross-level loops as one of its three sources.
- Dennett’s real patterns (LIT-220) and THEORY-036 bear on whether a psychological description that implementation keeps is a real pattern. This account does not need that question answered.