tag:blogger.com,1999:blog-5275657281509261156.post6891007206462113903..comments2024-03-28T04:04:55.806-07:00Comments on Faculty of Language: Physics envy bluesNorberthttp://www.blogger.com/profile/15701059232144474269noreply@blogger.comBlogger9125tag:blogger.com,1999:blog-5275657281509261156.post-85967795272413001712018-10-14T16:26:04.346-07:002018-10-14T16:26:04.346-07:00QM was very flexible in the beginning. But QM deal...QM was very flexible in the beginning. But QM deals with dicrete particles, with fields having descrete sets of states. It's a very clear game, and electron is an electron anywhere.<br /><br />But is a verb a verb anywhere? There is nothing comparable in, say, biology.Daniel N.https://www.blogger.com/profile/14585410511935134909noreply@blogger.comtag:blogger.com,1999:blog-5275657281509261156.post-62618848307915400592018-06-27T08:15:20.818-07:002018-06-27T08:15:20.818-07:00Insofar as there ARE "general parsimony"...Insofar as there ARE "general parsimony" arguments (well, they exist, certainly; the question is rather what force do/should they carry), following Elliot Sober's "Ockham's Razors" book, one might be able to assimilate as least at least one aspect of 'rigidity' to one approach to parsimony. Here are some quotes from reviews of the Sober book suggest why that might be so (emphases added by me)<br /><br />"According to the second paradigm, parsimony is relevant to estimating a model’s predictive accuracy since models with FEWER ADJUSTABLE PARAMETERS are less prone to “overfitting,” that is, less prone to describe random noise in the data rather than the true relationship of the variables of interest."<br />(Bengt Autzen, Philosophy of Science, v 83, 2016)<br /><br />"...when it comes to estimating the predictive accuracy of a model from a frequentist perspective, likelihood is only one relevant measure, another one being the NUMBER OF ADJUSTABLE PARAMETERS in the model."<br />(Michael Baumgartner, Australasian Journal of Philosophy, v 96, 2018)<br /><br />"...specify a metric for assessing the comparative parsimony of hypotheses of a particular type (e.g., NUMBER OF ADJUSTABLE PARAMETERS,..."<br />(Daniel Steel, Notre Dame Philosophical Reviews, 2016.1.27)<br /><br />IF having scarcity of adjustable parameters is a mark of rigidity (see John's comparison of general relativity with classical mechanics in his first comment above), then to the extent that Sober's views are correct, one might assimilate (some of) rigidity to (one aspect of one type of) parsimony. Perhaps. (I'm not a hedgehog--I like to share my hedges.)<br /><br />--RCRob Chametzkyhttps://www.blogger.com/profile/04943531685307739334noreply@blogger.comtag:blogger.com,1999:blog-5275657281509261156.post-11164968187972907522018-06-26T10:57:59.696-07:002018-06-26T10:57:59.696-07:00It looks great. Steven Weinberg's Dreams of a ...It looks great. Steven Weinberg's Dreams of a Final Theory is a classic in the same vein, although he is quite optimistic, as I recall.Anonymoushttps://www.blogger.com/profile/06409248369107264434noreply@blogger.comtag:blogger.com,1999:blog-5275657281509261156.post-73530655035041852582018-06-26T10:07:10.070-07:002018-06-26T10:07:10.070-07:00I haven't read Lost in Math yet though I have ...I haven't read Lost in Math yet though I have bought it -- maybe that will make it all clear. Thanks anyway.Alex Clarkhttps://www.blogger.com/profile/04634767958690153584noreply@blogger.comtag:blogger.com,1999:blog-5275657281509261156.post-18173209651103004242018-06-26T09:08:04.453-07:002018-06-26T09:08:04.453-07:00I don't think so. The idea is that binarity do...I don't think so. The idea is that binarity does explain phenomena in a way alternatives don't. So, binarity leads to the positing of covert structure, which explains novel phenomena you didn't have in mind at the time. It also insists that relations are binary - obviously - so you get to reduce c-command. Etc. Etc. Proliferating operations might fit the facts but produces shallow explanations close to the phenomena in the way a single operation dosesn't. I agree with Norbert's point that Occam offers little when it comes to deciding between alternatives.Anonymoushttps://www.blogger.com/profile/06409248369107264434noreply@blogger.comtag:blogger.com,1999:blog-5275657281509261156.post-28793054431075235352018-06-26T07:40:58.958-07:002018-06-26T07:40:58.958-07:00Isn't that more of a general parsimony argumen...Isn't that more of a general parsimony argument? I took Norbert's rigidity to be when a theory makes a bunch of predictions that can't be altered. I can't think of any very firm predictions that binarity makes. Alex Clarkhttps://www.blogger.com/profile/04634767958690153584noreply@blogger.comtag:blogger.com,1999:blog-5275657281509261156.post-71097930564695468152018-06-26T06:00:51.887-07:002018-06-26T06:00:51.887-07:00As regards binarity of merge, I had in mind the th...As regards binarity of merge, I had in mind the thought that if you want a single operation, only binary merge is possibly sufficient. Why? Let’s say you come across some phenomena that appear to require ternary merge (co-ordination, etc.). If you accept the appearances, then you lose the sufficiency of the single operation, for you can’t use ternary merge for [NP black dog], etc. So, you know that binary merge will suffice for a whole range of structures, and nothing else will do. Equally, showing how binary merge can handle co-ordination, double objects, etc. counts as a deeper explanation than positing two different operations, for the disparate structures at the surface are shown to follow from the single principle that is now proving to be actually sufficient. Of course, this thought trades on a bunch of empirical claims. My observation, here, only concerns the logic of the situation. So, the MP-ish/rigid hypothesis is that a single structure building operation suffices for all structures. (We know we can curry everything to the unary case, but that is a trick insofar as the new unary function will record the original arity.) <br /><br />One might think, ‘Well, why not n-ary merge as a basic principle? This is bound to work.’ This proposal introduces massive redundancy: for any structure we find, it could have been put together in a ‘flat’ way according to our n-ary merge. The explanation for why the structure is not flat, therefore, is shifted from merge to some other factors, such processing ones, perhaps. Alternatively, you could let some portion, as it were, of n-ary merge be used in this language or that, but that makes the merge hypothesis hyper flaccid. <br /><br />I should say, I think appeals to physics are interesting for paradigms of logic and explanation, but I don’t think the above reasoning relies upon any ‘third factor’ ideas. <br />Anonymoushttps://www.blogger.com/profile/06409248369107264434noreply@blogger.comtag:blogger.com,1999:blog-5275657281509261156.post-90554406715918258852018-06-26T04:13:19.050-07:002018-06-26T04:13:19.050-07:00Hi John,
Could you amplify your point about binar...Hi John, <br />Could you amplify your point about binarity of merge and how it relates to rigidity? Alex Clarkhttps://www.blogger.com/profile/04634767958690153584noreply@blogger.comtag:blogger.com,1999:blog-5275657281509261156.post-62819898859914265282018-06-26T02:03:41.784-07:002018-06-26T02:03:41.784-07:00A very nice discussion! The idea that rigidity is ...A very nice discussion! The idea that rigidity is a theoretical virtue goes back to Einstein. In his hands it means something like: a theory is right or wrong, but not modifiable without becoming nonsense. So, GR is rigid because it entails that gravitational attraction is inversely proportional to the square of the distance, and it can’t be modified to entail that it is the cube of the distance, to accommodate, say, a possible world where gravity is measuable in terms of the cube of the distance. GR is unlike classical mechanics, in this respect. So, right, rigidity is good, because it explains phenomena in a very deep way, as virtual physical necessities (MP-ish pun intended). If GR is right at all in any respect, then gravity MUST be a function of the square of the distance. This kind of reasoning applies in spades to linguistics, such as issues of structural dependence vis-à-vis PoS issues, the binarity of merge, etc. If one thinks there is nothing deep going on, little wonder that the explanations on offer are shallow. <br />Anonymoushttps://www.blogger.com/profile/06409248369107264434noreply@blogger.com