Loading ...
Sorry, an error occurred while loading the content.

14020Re: [ANE-2] Re: Oldest writing system

Expand Messages
  • Peter T. Daniels
    Jan 6, 2012
    • 0 Attachment
      Is this an ANE topic?
       
      Rao's "claims about statistical evidence for the linguistic nature of the Indus symbols" are self-evidently, and evident to apparently all but the editors who published them, utterly irrelevant to the study of Indus writing and hardly worth detailed refutation.
       
      Richard was, then, at BLS apparently donning straw arms to combat a straw man. I'm not surprised that nearly two years later the only impression I have of the talk is as I said.

      The evidence for Dravidian being the most likely of the possible languages for Indus writing is gathered in Parpola's 1994 book. Curiously, a recently-published Harvard dissertation, which was delayed for years by the intransigence of Witzel, argues that it can't possibly be Dravidian but most likely is Munda.

      --
      Peter T. Daniels grammatim@...
      Jersey City


      >________________________________
      > From: richardwsproat <rws@...>
      >To: ANE-2@yahoogroups.com
      >Sent: Friday, January 6, 2012 3:53 PM
      >Subject: [ANE-2] Re: Oldest writing system
      >
      >
      >

      >
      >I would add two further comments to Steve's comments here, both related to Daniel's misrepresentation of what I said in the Berkeley talk (and elsewhere).
      >
      >First of all, to understand the context of the point about heraldry that I was making, one needs to understand another set of papers not cited by Steve that deal with Rajesh Rao's claims about statistical evidence for the linguistic nature of the Indus symbols. My own discussion of this can be seen in a "Last Words" piece in Computational Linguistics (http://www.mitpressjournals.org/doi/abs/10.1162/coli_a_00011) and subsequent replies by Rao and colleagues (also Rob Lee and colleagues) and my reply to their reply. The essence of the claims was not that heraldry should be considered writing, which is obviously absurd, or that since heraldry isn't writing therefore the Indus symbols should not be. The point, which clearly Daniels did not understand, was that if one wants to take such statistical arguments seriously, then one also has to deal with the fact that many systems that we KNOW are not writing, "look like" writing from the point of view of those
      statistical methods. I can't speak for the Berkeley Linguistics Society audience, but the audience of computational linguists at the Empirical Methods in Natural Language Processing conference in Singapore the previous year clearly understood what I was arguing.
      >
      >Second, the BLS meeting that Daniels referred to also featured an invited talk by him. During that talk he mentioned the Indus symbols and, practically looking directly at me, emphasized that they "ARE Dravidian". The evidence for that conclusion is, quite frankly, pathetic, and the fact that Daniels believes that the evidence is good makes me wonder about the standards of evidence being applied.
      >
      >Richard Sproat
      >Center for Spoken Language Understanding
      >Oregon Health & Science University
      >
      >--- In ANE-2@yahoogroups.com, Steve Farmer <saf@...> wrote:
      >>
      >> Diana Gainer writes, in this rather silly thread on the "first writing system":
      >>
      >> >> The Harappan script may be another proto-writing
      >> >> system (Farmer, Sproat, and Witzel 2004 "The Collapse of the Indus Script Thesis: The Myth of a Literate Harappan Civilization").
      >>
      >>
      >> If you read the paper you'll find that we certainly don't make that claim -- quite the opposite. We offer evidence instead that these extremely short symbol chains -- we have many thousands of them on over a dozen different types of materials, all short -- were non-linguistic in nature, not "proto-writing." The older claim (by Parpola, etc. -- made e.g. in Daniels and Bright's studies of writing systems) was that they were supposedly part of a fully functioning writing system. That claim was made based on finds going back to the 1870s and wasn't questioned until 2004, when we published our first paper on this.
      >>
      >> Note now that Parpola has since backtracked and *now* claims -- pushed by the evidence that we've presented, as he noted in a conference in Japan we both attended two years ago -- that they were part of a so-called proto-writing system. But this isn't our argument. And it wasn't his either before we published our first papers.
      >>
      >> On what we say, see our 2004 paper, pp. 33 ff., where we specifically argued against the proto-writing argument, long before Parpola abandoned his claim that this was a "full writing system" (in Daniel's volume) and fell back to the proto-writing argument:
      >>
      >> http://www.safarmer.com/fsw2.pdf
      >>
      >> Peter Daniels writes, re. our (non-existent) claims that this is "proto-writing":
      >>
      >> > Please don't take Farmer/Sproat/Witzel seriously. Richard Sproat's only?/best? argument that Indus writing isn't writing is that the blazons used in heraldry would then have to be considered a writing system (presented at the 2010 Berkeley Linguistic Society meeting, not yet published), and since it isn't, therefore that shouldn't be. The claim was so absurd that no one in the audience could even figure out how to challenge it.
      >>
      >> Grotesque misreporting, showing that Daniels hasn't read our papers. This certainly is NOT Richard Sproat's argument.
      >>
      >> It helps to read our papers before you say anything, Peter. Start with our 2004 paper, which was covered in _Science_ magazine and has been downloaded many hundreds of thousands of times in reprints from one server alone (see above link). We discuss blazons in that paper, but certainly not the way you claim.
      >>
      >> Our data have been presented at a long string of conferences held at Harvard (2001, 2002, 2003, 2004), in Kyoto (2005, 2009), at Stanford University (special Linguistic Society of America conference on our work, 2007, attended by many writing specialists) at the U. of Bologna, and many other places.
      >>
      >> I would *love* to hear Daniel's arguments that a society that produced ludicrously short symbol chains over a 600 year period -- but left not one "text" behind, despite supposedly being fully literate (as Parpola argued in Daniel's 1996 volume, but no longer accepts himself) -- was a full writing (i.e., speech encoding) system.
      >>
      >> The old claim that the Indus DID write long texts but only on perishable materials is trivial to debunk: we don't know of any literate society anywhere in the world that supposedly wrote long texts on perishable materials for hundreds of years but never left a *single* "text" behind on durable materials.
      >>
      >> Steve Farmer

      [Non-text portions of this message have been removed]
    • Show all 24 messages in this topic