Abstract It is argued that John Bickle's Ruthless Reductionism is flawed as an ac count of the practice of neuroscience. Examples from genetics and linguistics suggest, first, that not every mind-brain link or gene-phenotype link qualifies as a reduction or as a complete explanation, and, second, that the higher (psychological) level of analy sis is not likely to disappear as neuroscience progresses. The most plausible picture of the evolving sciences of the mind-brain seems a patchwork of multiple connections and partial explanations, linking anatomy, mechanisms and functions across different domains, levels, and grain sizes. Bickle's claim that only the molecular level provides genuine explanations, and higher level concepts are just heuristics that will soon be redundant, is thus rejected. In addition, it is argued that Bickle's recasting of philos ophy of science as metascience explicating empirical practices, ignores an essential role for philosophy in reflecting upon criteria for reduction and explanation. Many interesting and complex issues remain to be investigated for the philosophy of science, and in particular the nature of interlevel links found in empirical research requires sophisticated philosophical analysis. (..) Bickle: claim that a direct bridge from behaviour to molecule has been found (example: social cognition memory consolidation in mice as related to molecular mechanisms in the hippocampus - disrupted in mutant (CREB) mice. (..) a link is not a [proven] reduction (..) knock-out parts does not explain the whole phenotypic trait (cf removing a carburator halts the car, but the carburator does not completely explain the work of the car-engine) (..) cells and molecules do not constitute a reduction of cognition but are just one part of a complex causal web. (..) what constitutes adequate reduction is a philosophical issue. (..) the picture of multilevel anarchistic level-hopping seems empircally adequate. To give one example: event related brain potentials (ERPs) can be used to find out about cognitive processes. (..) For example, syntactic and semantic processes in sentence comprehension can be functionally and anatomically differentiated, and the sequence and timing of syntactic and lexical stages in speech production can be traced (Hagoort 2003). (..) This kind of level-hopping seems well captured in HIT (McCauley & Bechtel 2001). (..) we probably cannot get the explanatory story (the equipotent image) of asso ciative memory from LTP; when functional and mechanical kinds and explanations cross-classify, a purely bottom-up explanation will not be forthcoming. (..) the functional and the motor neuron level of behaviour are in important ways distinct, and ruthless reductionism has a long way to go before the explanation of mice social behaviour in terms of a causal chain from hippocampus to motor neuron makes the functional description otiose. (..) Feyerabend's notorious "anarchism" is really pluralism (Tsou 2003, p 210, note 3), emphasising proliferation of theories, and is motivated by interest in and respect for scientific progress. (..) Feyerabend (1968) argues that the logical empiricists hamper progress by imposing methodological conditions (consistency and meaning invariance) on theory dynam ics. The antidote to these unproductive constraints is variety. (..) Let's forget the single-level fixation, and be careful about simplistic interpretations of interlevel relations. |