ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Semantic Dementia

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Wed, 20 Nov 2013 17:54:58 -0500
Message-id: <528D3DC2.4030205@xxxxxxxxxxx>
On 11/20/2013 2:33 PM, Rich Cooper wrote:
> I find their hub-and-spoke model very interesting when trying
> to understand how the brain uses language, sensors and effectors
> in such flexible ways:    (01)

I completely agree.  But I don't believe that there is a single hub,
and I don't believe that the authors are claiming a single hub.    (02)

And guess what you get when you have multiple hubs?  A network.    (03)

The linguist Sydney Lamb, who had been collaborating with
neuroscientists since the 1970s, formulated his theory of
neurocognitive networks.  I summarize Lamb's theory (with
references) in http://www.jfsowa.com/talks/goal2.pdf    (04)

I adapted (with permission) several of Lamb's diagrams in
those slides.  In particular, see slides 25 to 29.    (05)

 From the paper you cited (by Hoffman et al.)
> This model holds that in addition to modality-specific sources
> of information ("spokes") and their inter-connections, representation
> of conceptual knowledge requires an integrative "hub". The hub
> uses information from the modality-specific spoke regions to develop
> modality-invariant, conceptual representations that capture deeper
> patterns of conceptual similarity across all sensory-motor and
> verbal modalities.    (06)

I certainly agree with that point, and so would Syd.    (07)

In slide 28 goal2.pdf, Lamb had a "hub" marked C (for 'concept')
in the parietal lobes.  That region of the brain is specialized for
cognitive maps or patterns that relate spatial relations to the
sensory-motor modalities.  The parietal lobes are critical for
linking various sensory-motor regions with spatial and auditory
regions.    (08)

> The hub-and-spoke framework holds that the ATL hub provides this
> critical aspect of conceptual representation through the formation
> of representations that integrate information from all sensory-motor-
> verbal domains.    (09)

Syd and I would agree that the ATL hub (Anterior Frontal Lobe) is
critical for certain kinds of information, which could be called
a "critical aspect" of conceptual information.    (010)

The question of which, if any, of the many nodes in the cortex should
be designated as the principal node for a concept is open for debate.
The most likely answer is that many different hubs are specialized
for different aspects of what we might call conceptual information.    (011)

Note that "conceptual" information about verbs is located in BA 47
of the frontal lobes, just below Broca's area.  See slides 15 to 18
and slides 23 to 27.    (012)

In any case, this article is an important contribution, and I plan
to add another slide to goal2.pdf to cite it.    (013)

John    (014)

PS:  Sydney Lamb was the thesis adviser to George Heidorn, who went to
IBM Research.  One of the early implementations of conceptual graphs
used a version of the software based on Heidorn's dissertation.  See    (015)

    http://www.jfsowa.com/pubs/cg1986.pdf    (016)

George H. and Karen Jensen later went to Microsoft Research, where they
implemented the grammar checker for MS Word.  If you saw any little
green squiggles, you can thank George, Karen, and Sydney.    (017)

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (018)

<Prev in Thread] Current Thread [Next in Thread>