- From: Pat Hayes <phayes@ihmc.us>
- Date: Fri, 24 Mar 2006 11:11:44 -0800
- To: "Steve Pepper" <pepper@ontopia.net>
- Cc: "SWBPD list" <public-swbp-wg@w3.org>
>Hi Pat, > >Your intervention in this discussion is much appreciated. See if you still agree at the end of this reply :-) >| >[Regarding the issue of whether reification has the same >| >semantics in Topic Maps and RDF] >| >| Sorry, I have only just seen this thread. > >This reply doesn't attempt to answer all your questions: others >are better qualified to do that. But a couple of comments might >help the discussion along. > >1. The basic fact that Bill and Sue are married (your #2) can >be represented in TMs as a typed binary association between >what TMs term the two "subjects" Bill and Sue who play the >roles of husband and wife. Thus (in LTM notation, which is >hopefully self-explanatory): > > married-to( bill : husband, sue : wife ) > >This association can be reified such that the relationship it >represents itself becomes a subject which can play roles in >further relationships: > > married-to( bill : husband, sue : wife ) > ~marriage-of-bill-and-sue > >Thus it is possible (in TMs) to state that (A) Bill and Sue are >married and that (B) their marriage dates from a certain date, >as two separate associations and yet still know that we are >talking about the same "thing" (the relationship, in the common >use sense of the word): OK, that makes sense. So these are tropes, the particular holdings of a relation between some number of individuals, sense 3. (BTW, there are many other equally common uses of 'relationship', so to appeal to a common usage without further explanation is kind of sloppy. But I digress.) In a reply to Lars' message I sketched a way to map this into RDF, based on this intuition. > > (A) married-to( bill : husband, sue : wife ) > ~marriage-of-bill-and-sue > > (B) {marriage-of-bill-and-sue, start-date, [[2005-08-20]]} > >Note: (B) is here expressed as an occurrence since the date is >represented as a string rather than a topic. I completely missed that last point (a string is not a topic? Never mind...) but I hope this doesn't matter. The main point is that marriage-of-bill-and-sue is a thing that can have properties, such as a starting date. >| Then we can distinguish >| >| 2. The fact that Bill and Sue are married >| 3. The particular state of being married that >| holds uniquely between Bill and Sue (and no >| others) > >2. As Lars Marius pointed out, it is hard to see the difference >between these two from the Topic Maps point of view. It is not really a matter of point of view, simply of clear thinking. This article, for example, http://plato.stanford.edu/entries/tropes/ discusses the issues very thoroughly. > Perhaps >this is precisely because Topic Maps is NOT based on logic, but >rather seeks to model a fuzzier and more natural language >oriented view of the world? I will try to avoid giving offense, but I suspect it has more to do with the fact that the foundations of TM have not yet been adequately developed to bring them up to current standards of scholarship in this area. One can aim for precision in a theory of fuzziness or ambiguity of meaning, and indeed such theories are now many years mature. And there is no 'natural language oriented view of the world': the precise study of linguistic meaning uses the same formal tools that are used in logical semantics. Semantics is just the study of how notations express things about a world, and formal semantics is just semantics done with precision, i.e. using mathematical techniques. Do not make the mistake of assuming that because natural language is 'informal', therefore it cannot be studied using precise theories. It can, and is. Anyway, thanks for your explanation, which was very helpful. BTW, I bet one could write a formal semantics, i.e. a model theory, for TM fairly straightforwardly. Would this be of interest or utility to anyone? Pat > >Steve > >-- >Steve Pepper <pepper@ontopia.net> >Coordinator, W3C RDF/TM Task Force -- --------------------------------------------------------------------- IHMC (850)434 8903 or (650)494 3973 home 40 South Alcaniz St. (850)202 4416 office Pensacola (850)202 4440 fax FL 32502 (850)291 0667 cell phayesAT-SIGNihmc.us http://www.ihmc.us/users/phayes
Received on Friday, 24 March 2006 19:11:55 UTC