This is a preview of subscription content, log in via an institution to check access.
Notes
There is a strand in discussions of emergence that points in that direction—discussions linking emergence to unpredictability, but I think that these are explicable in other ways. Lyapunov divergence in dynamical systems—associated with discussions of chaos—yields practical unpredictability (for anything short of knowing the values of variables to infinite precision) with deterministic systems that are totally understood. And discussions of unpredictability sometimes refer to the qualitative newness of the emergent property—which however is not at all inconsistent with its after the fact explicability, or even before the fact predictability, given a description of the initial conditions and mechanistic characteristics of the system.
I’ll discuss below why he might mistakenly regard my position as epistemological—I believe that he thinks that my sort of pragmatism, being heuristic, must give up on any metaphysical claims. On this we disagree.
One might ask why Rosenberg or most metaphysicians think that reducability has a bearing on metaphysics. This is because they suppose that if something is reducable, it is ontologically derivative, making it in principle eliminable, and thus somehow less important or less real than the things it is derivative from. It doesn’t have to be included in the basic ontiology. Given that I root reality in robustness (see the next section) I do not accept this view. Even if something were totally derivable from something else, I do not accept that as compromising its reality if it is robust. But furthermore, in so called inter-level reductions, the derivational relations involve approximations, and are not deductive so the reality of the upper-level things—generated by their multiple-detectability at the upper level—are not even on Rosenberg or Kim’s views eliminable. (Replaceable perhaps, but not deductively eliminable.)
When I feared that there might be virtually no cases, I was delighted to realize that the properties that were the subjects of the great conservation laws of physics were aggregative functions of their parts. Indeed, that’s why energy, mass and charge are conserved across all sorts of transformations! This was an unexpected (but welcome) result. Nature fits!
I don’t reject either label as long as they are not taken as implying anti-realism, as they commonly are. I count myself a realist.
One of the most intriguing is that it commits me to admit degrees of reality. Things may be detectable in different numbers of ways, each of different intrinsic reliabilities, and which can fail under different combinations of conditions. It also looks as if it is defeasible: isn’t it possible that the different independent ways could all happen to fail under the same conditions? And how do we demonstrate independence, and in what respects? I have already discussed failures of independence as “pseudo-robustness” and there is no reason why claims of robustness cannot be defeated. There are no magic bullets in science, and robustness isn’t one either, but it is the best that we have got in many circumstances, and is widely used in science and in mathematics for that reason. Interest in robustness has been growing in recent years (see Soler et al. 2011; Wagner 2005).
Perhaps we make more progress here if we ask not which is correct, but ask which one do we use to correct the other and when.
Indeed, in his classic ground-breaking paper, von Neumann (1956) considered the use of k-out-of-m choice “organs” to deal with the unreliability of computer components in large computers. Thus, here a (2,3) “majority” organ would fire at t + 1 if 2 out of 3 of its input lines fired at t. Computer reliability was a serious problem in early days, when Boolean states were represented by vacuum tube diodes. Given that electronic tubes burned out not infrequently, it was thought that there was a (relatively small) maximum size for a computer if the program was to be loaded and run before one of the tubes burned out. (Early in the development of computers this was estimated to be as small as 1,200 tubes.) With IC’s now having upwards of 50 million components and software developments for parallel processing and detecting and parsing out failed components, we have made a lot of progress.
References
Bedau M, Humphreys P (eds) (2008) Emergence: contemporary readings in philosophy and science. MIT Press, Cambridge
Campbell D, Fiske D (1959) Convergent and discriminant validation by the multitrait-multimethod matrix. Psychol Bull 56(2):81–105
Ruse M (2008) Handmaiden to the science? Am Sci 96:340–342
Soler L et al (eds) (2011) Robustness of science. Springer, Berlin (forthcoming)
Trout JD (1998) Measuring the intentional world. Oxford University Press, USA
Von Neumann J (1956) Probabilistic logic and the synthesis of reliable organisms from unreliable components. In: Shannon CE, McCarthy J (eds) Automata studies. Princeton University Press, Princeton, pp 43–98
Wagner A (2005) Robustness and evolvability in living systems. Princeton University Press, Princeton
Wimsatt W (2007) Re-engineering philosophy for limited beings: piecewise approximations to reality. Harvard
Woodward J (2005) Making things happen: a theory of causal relations. Oxford University Press, USA
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Wimsatt, W.C. Robust re-engineering: a philosophical account?. Biol Philos 26, 295–303 (2011). https://doi.org/10.1007/s10539-011-9260-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10539-011-9260-8