CAS LX 502
Semantics
2b. Sense and concepts
2.4-
Mental representations
• We can talk about things that don’t exist, things as
they might have been, things as they will be.
Language is not limited to references to tangible
entities and properties in the real world.
• We have a mental representation of the meaning
of language (and much of what we are studying
here are the properties of these mental
representations).
*Images
• The mental representation of the meaning of
a word, even a name, cannot simply be an
image or other perceptually grounded
representation (justice? happiness?). Rather,
we have a complex mental concept as the
meanings of words—their sense.
• Perceptual representations may be part of the
concept, but they are not the all there is to the
concept.
Concepts transcend words
• Of course, there
need not be a word
for there to be a
concept. Concepts
can be described,
and given a
designation if they
are useful.
sniglet: n. Any word that doesn’t
appear in the dictionary but
should.
• Rich Hall had a series of books
in the 80’s containing these.
eastroturf: n. The artificial grass
in Easter baskets.
furnidents: n. The indentations
that appear in carpets after a
piece of furniture has been
removed.
What is our concept of bird?
• When we think of the meaning of bird, the
concept of bird, we think of it as applying to some
things and not others. At a first pass, it might have
to satisfy these necessary conditions (linking this
concept to other concepts: animal, wings, …)
•
•
•
•
•
It is an animal
It has two wings
It can fly
It sings or chirps
…
Necessary and sufficient
• Not all of these linked concepts are really
necessary for something to fall under the
concept bird.
• Is a penguin a bird? Surely.
Yet it doesn’t fly. Is a fish with three
eyes really a fish?
So what are the
sufficient conditions?
According to Webster
• Bird: n. 2 : any of a class (Aves) of warm-blooded
vertebrates distinguished by having the body more
or less completely covered with feathers and the
forelimbs modified as wings.
• Wing: n. 1 a : one of the movable feathered or
membranous paired appendages by means of
which a bird, bat, or insect is able to fly; also :
such an appendage even though rudimentary if
possessed by an animal belonging to a group
characterized by the power of flight
According to Webster
• Bird: n. 2 : any of a class (Aves) of warm-blooded
vertebrates distinguished by having the body more
In other
or less completely covered with feathers and the
words, we ask
forelimbs modified as wings.
a bird expert.
• Wing: n. 1 a : one of the movable feathered or
membranous paired appendages by means of
which a bird, bat, or insect is able to fly; also :
such an appendage even though rudimentary if
possessed by an animal belonging to a group
characterized by the power of flight
Thus: penguins
are birds?
Folk semantics
• Yet, we’re pretty happy to call penguins birds, as
well as ostriches. We might call a whale or a
dolphin a fish, or we might call a tomato a
vegetable, even if the experts would tell us
otherwise. We may be able to discern the
difference between an elm and a beech only by
seeing how an expert reacts when faced with one.
• We have some kind of concept of fish, vegetable,
bird, elm that doesn’t seem to rely on a strict
definition (necessary and sufficient conditions).
Prototypes
• So how is our conceptual knowledge organized?
Perhaps a concept has a prototypical member,
meeting all of the conditions, where things that
meet only some of the conditions (having only
some of the characteristic features) are more
peripheral. The closer the match, the more typical.
• Some features are clearly more fundamental than
others. A mechanical bird would seem a much less
typical bird than an ostrich—probably not even a bird at
all.
Characteristic features,
prototypical exemplars
• In a sense, these ideas make intuitive
sense—but in another sense, they are no use
to us. These don’t seem to tell us anything
about how concepts can and can’t be
organized, really. They don’t explain why
concepts are the way they are. They just
give us a language of description once we
already know the facts.
The acquisition of concepts
• Any attempt to define a concept will quite rapidly
run into a circularity problem. Concepts are
clearly related to one another.
• Cf. conceptual networks: A duck is a bird, and inherits
from bird the conceptual associations bird has, such as
being an animal, inheriting the conceptual associations
animal has, etc.).
• But eventually, there must be nothing to reduce to.
(Or is it turtles all the way down?)
The acquisition of concepts
• This is a problem for the philosophers, perhaps,
but it does not seem beyond the realm of
possibility that there are a certain number of
fundamental concepts that we start out with
(perhaps like the structure of language attributable
to UG), and from which other concepts are
derived. The philosophical implications are many.
But we do seem, as kids, to know how to
characterize doggie when a doggie is pointed out
to us (cf. Quine’s gavagai).
Language, thought, and reality
• “Don’t you see that the whole aim of Newspeak is
to narrow the range of thought? In the end we
shall make thoughtcrime literally impossible,
because there will be no words in which to express
it. Every concept that can ever be needed, will be
expressed by exactly one word, with its meaning
rigidly defined and all its subsidiary meanings
rubbed out and forgotten.” (according to Syme,
from George Orwell’s 1984, of course)
Sapir-Whorf
• B.L. Whorf, studying Uto-Aztecan languages,
observed that agreement systems, declensions,
tense-marking vary widely across languages, and
languages often divide word into classes based on
rather arbitrary criteria.
• He has since been much maligned for suggesting
that these classifications impose a restriction on
how speakers classify concepts—taken to an
extreme, this leads to the idea of Newspeak.
Known as the Sapir-Whorf hypothesis.
Pinker v. Whorf
• Pinker (The Language Instinct) runs through a
number of arguments against the idea that
language determines concepts.
• 5-mo olds: react to unexpected changes in cardinality.
• Vervet monkey’s sister: bit VM’s antagonist’s sister.
• Mental rotation: 56 RPM.
• So what is the language of concepts? Not English.
We might call it Mentalese, if we so wish, but
whatever we call it, it is richer, more
unambiguous, and certainly distinct from spoken
languages.
 snow
 powder
 flurry
 sleet
 hail
 blizzard
 dusting
 hardpack
 slush
 avalanche
Descargar

CAS LX 502 - Boston University