home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!auvm!FAC.ANU.EDU.AU!ANDALING
- Message-ID: <9211192255.AA24857@fac.anu.edu.au>
- Newsgroups: bit.listserv.csg-l
- Date: Fri, 20 Nov 1992 09:55:00 EST
- Sender: "Control Systems Group Network (CSGnet)" <CSG-L@UIUCVMD.BITNET>
- From: Avery Andrews <andaling@FAC.ANU.EDU.AU>
- Subject: images, etc.
- Lines: 40
-
- [Avery Andrews 921129.0930]
-
- So the way it looks to me at this moment, the perceptual reference
- `signals' that you would derive from the text to tell you where to go
- would actually have a lot in common with the `mental models' postulated
- by people like Johnson-Laird. Their supposed properties are (to my
- mind at least) a rather unclear combination of imagistic & and logical
- properties, & which Bill's view of things might help to clarify.
-
- So next question: how do we distinguish between:
-
- a picture of a cat chasing a dog
-
- a picture of a dog chasing a cat
-
- Maybe bonkers idea. What if there were two areas, an event area where
- patterns of neural activity corresponding to verbs occurred, and an
- entity, where patterns corresponding to the entities occurred, with the
- normal situation being that the event pattern (thought of as a static
- configuration of some fibers firing and others not) coexists with
- a sequence of entity patterns, in a standard order (presumably
- Agent - Patient, though there's lots of other roles to fit in).
-
- These would get correlated with perception via complex events being
- scanned in a characteristic order. E.g., `chase' refers to a perception
- whereby on the entity level, you look first at the one behind and
- register its features on the entity level, and then do so for the
- chasee, while `flee' scans in the reverse order.
-
- Remembering a complex perception would then amount to remembering a
- sequence of scanning activities & attendent results. How this is to
- be achieved I don't have any clear ideas. But, for what it's worth,
- the inner representations I'm suggesting have pretty much the same
- organization as appears overtly in manual sign language, where
- signs representing actions are articulated simultaneously with those
- designating participants (coded as pointing gestures to positions
- in space, which have been previously been associated with specific
- referents).
-
- Avery.Andrews@anu.edu.au
-