EFTA00756622.pdf

From: Martin Nowak To: Jeffrey Epstein <jeevacation®gmail.com>, Corina Tamita Subject: Fwd: [forgot to cc] Date: Sun, 12 Sep 2010 19:32:31 +0000 Attachments: GeorgeWilliams.doc Begin forwarded message: From: Date: September 12, 2010 3:06:09 PM EDT To: "Martin Nowak" •cr > Subject: [forgot to cc] Original Message Subject: viel in Dank From: Date: Sun, September 12, 2010 3:03 pm To: "Karl Sigmund" lieber karl viel in herzlichen Dank fuer ihren e-mail. i was glad you liked my private letter better than the public since i have already withdrawn my name from it--i do not believe in attacking someone twice, certainly not in private and then again in public. also it has now grown in size and co-authors such that i am biased in the opposite direction, i.e. if i see 117 people running in one direction, i naturally assume the opposite one has something to do say for itself! "In contrast to Martin, I think that indirect fitness is a useful concept, and that inclusive fitness can, in principle at least, be well defined even without all the assumptions about weak selection etc. Like with the Price equation, we are talking about a method of book-keeping. It is a matter of taste whether one finds it enlightening or not. But it is certainly possible to use alternative methods of accounting to track selection-mutation effects. It may make sense to compute the inclusive fitness of a worker ant, but it probably stops making sense to compute the inclusive fitness of a body cell in a multicellular organism." I AGREE WITH ALL OF THIS "However, a school of thought (Alan Grafen, Stu West, Andy Gardner, Peter Taylor, ...) has elevated the IF method into a mantra. Their actual results are quite limited, EFTA00756622 circling endlessly around the island model, assuming weak selection etc. They lead to some interesting results, but this does not imply that IF is THE true and only explanation of all forms of cooperation. This claim is implied in a lot of their work, however. It does not get better when Gardner and West, for instance, pontify that 'what should no longer be done' is using game theoretical models, as that infamous Nowak has done (they did not mention me, by the way, nor you, Bob: but certainly, reciprocal altruism is not included in their list of 'what should be done')." I AGREE WITH ALL OF THIS AND WAS IGNORANT OF IT UNTIL YOU WROTE. i do not bother to keep up with literatures on subjects i think i understand, if you know what i mean, but a friend backed you up completely (and filled in some details over the phone). From his e-mail: "Karl is right on and describes the school of hegemonic inclusive fitness quite well.." So, vielen Dank for that. "Quite generally, such trifling things as frequency dependent selection, between-species mutualism, multi-level selection etc. do not seem to rank high on their agenda. I should add that while I think that non-random assortment is all-important for the evolution of cooperation, to reduce it all to relatedness is to engage in semantic break-dance." ABSOULUTELY--AND I HAD NOTED THAT BEFORE YOUR LETTER, AN EFFORT TO SWALLOW EVERYTHING INTO KINSHIP WHEN THIS IS MANIFESTLY FALSE. By the way, when i first tried to model reciprocal altruism, i had genes at different unlinked loci dictating whether i cooperated with you and you with me--precisely to prevent the problem of identical genes helping themselves in others being a hidden part of the argument--but it was too complicated for me to develop. "I think that relatedness should be used as the 'man on the street' uses that term, and not be replaced by increasingly complicated statistical expressions. I am ready to acknowledge any rule 'b/c>something' as an extended version of Hamilton's rule, but do not think it is fruitful to baptize that 'something' as relatedness. I can conceive of some 'Hamilton's rule' for cooperation between species, based on preferential assortment. But hailing it as kin selection would be going too far." COULD NOT AGREE MORE; first time i have seen 'something' substituted for 'r'--ganz qvatch once again--there are an endless number of 'somethings' that could interact with the cost/benefit ratio of an act so as to produce coherent patterns: Jesus Christ, Karl, aber wass geht doch hier? "I should have thought that it is enough to quietly wait till the current infatuation with that new, very technical, unnecessarily complex and, in my eyes, contrived version of IF fades down. But if there is one thing Martin cannot do, it is waiting. I see his paper as a reaction to a provocation." EFTA00756623 I SEE YOUR POINT "Of course a good strategist should never get provoked, but it may actually be beneficial to the health of the field that the furor erupts sooner rather than later." I GOT A LAUGH OUT OF THIS ONE; first i agree--given your characterization of what has been going on--and second it applies to me as well i shouldn't let martin's past possible misbehavior get the best of me but perhaps it could be bracing for him nonetheless; if he substitutes the kinds of careful papers on cooperation that i have admired for some 20+ years with this kind of performance, very weak on haplodiploidy and eusociality, loaded with mis-assertions both empirical and theoretical, and then weds it to a poorly focused and misleading treatment of alternative forms of stating inclusive fitness, he will only bring himself down; we both know he is ambitious as hell, which perhaps partly explains his sensitivity to provocation--aber bitte doch martin, DENKEN, DENKEN, DENKEN--that is what is in short supply in academia, not models or grand assertions but careful thought; you have a talent, use it i have two last, little points: I --the only thing that ever bothered be about hamiltons's version was that as an actor you had to strip yourself of the benefits of relatives on you and call THAT your fitness or RS--but that is, virtually by definition, unmeasurable in nature i see that problem surface in nowak et al but perhaps missed any advance on clarifying the different accounting systems--e.g. do we leave out kin effects on our phenotype and add to this our effects on their phenotypes or vice-versa, people tell me they are logically equivalent, but i do not know 2--the acid test of any kinship formulation for me is genomic imprinting, itself the only advance in theory since hamilton worth mentioning. For example, you can show that inclusive fitness language can be made isomorphic to group selection (within and between group) language but it would take you 300 years to predict genomic imprinting in the latter system but took no time at all in the 'what is the probability of an identical copy of this gene in another individual' system of language (and logic) first invented by bill so, if someone wants to rewrite the mathematics of kinship let them always ask--does this help or hurt my ability to predict genomic imprinting? alles gute von nuevo jersey and vielen Dank EFTA00756624 robert p.s. i attach a remembrance of George Williams EFTA00756625