Re: Questions 11, 14, 24, 28, 32 & 35 Challenged

From: Harnad, Stevan (
Date: Wed May 21 1997 - 18:25:54 BST

> From: Linketscher, Nadja <>
    (11) The difference between sentence meaning and speaker meaning is:
    a. ***one is semantic and the other is pragmatic
    b. one is sense and the other is reference
    c. one is innate and the other is learned
    d. one is semantic and the other is syntactic
    e. one is syntactic and the other is pragmatic

> I thought e is the right answer. As far as I understood it,
> sentence meaning=literal meaning which is 'invariant over
> different contexts of use' (glossary in Green, p.375). Does
> this not imply that it is context independent? And taken
> into account that syntactic rules apply to the sentence
> meaning only, the answer should be, I think, e.

Context-independent does not mean syntactic. Syntax is a set of rules
for manipulating symbols on the basis of their (arbitrary) shapes rather
than their meaning. Context-independent sentence-meaning just means
the literal meaning (not shape, not syntax) of an utterance (sentence).

Both sentence meaning and speaker meaning are "meaning." The difference
is that the first is context-independent and the second is not. The
second is based on pragmatics, the first only on semantics.

    (14) Which of the following could be understood from sentence meaning
    a. he gave her the book
    b. he is a personal friend
    c. ***"he" is a personal pronoun
    d. this is where it happened
    e. I can't understand this

> Here, the answer seems to depend on the right answer in the
> question raised above. Even though you do not know what the
> person did not understand in answer e, it seems obvious and
> not requiring context that there is a communication problem
> which needs further clarification.

Yes, but what is "this"? You had to be there to know; or you need
a context of more sentences. On its own, "this", and hence the sentence,
cannot be understood: The is true of any sentence that has pronouns (he,
she, it, we, they, you) or deictic terms (this, that, here, there).

The only exception is sentence c -- and that is because "he" is not
being used to refer to anything; it's just a piece of vocabular that
the sentence, all by itself, is referring to: It's all there in the
sentence meaning.

    (24) Why is "linguistic theft" a victimless crime?
    a. because people can steal your words but not your ideas
    b. because if you copy someone else's answers, that person doesn't lose
    c. because you only need to say as much as you want to
    d. a and c
    e. ***none of the above

> What is wrong with answer d? Take an example: say, you have
> developed a theory, which can be simulated by computation,
> (so you actually explain something),and therefore you have
> made an influential contribution to cognitive science. When
> explaining the underlying mechanisms in PY 104, you do not
> have to tell us all about it, you just say as much as you
> want to (c). And, even if you did tell us anything about it,
> and someone rephrases it in his or her own words,it is still
> your idea (a, similar to b) - so, what do you lose?

Well, a is wrong because it's irrelevant to this meaning of linguistic
"theft" (and because it's not even true: people CAN steal your ideas!).

b is wrong, because it's also irrelevant (although this one did have
the FLAVOUR of the correct answer, and could have been misinterpreted as
the right answer by someone who knew everything that needed to be known on
this problem; see below).

And c is also wrong because irrelevant: The correct answer would have
been: Language allows you to learn things by having them explained in
words by someone who already knows. It saves you from having to learn
many things from experience through trial and error. Except in special
cases -- where the topic of the sentence is something you and the speaker
both want, and telling you would put the speaker at risk of losing the
prize to you -- the hearer gains while the teller loses nothing.
Here was the background information from my lecture on Chapter 9:
    If you had to eat mushrooms to survive because there was nothing
    else to eat, nibbling a little and waiting to see whether or not
    you got sick would definitely be the hard way of getting your
    knowledge about mushrooms. Let's call that way of getting your
    knowledge -- by trial and error experience, guided by feedback from
    the consequences of what you do -- "honest toil."
    Language makes it possible for you to get the same information by a
    kind of "theft": If someone who knows the features of the edible
    and poisonous mushrooms describes them to you in words, then you
    have saved yourself a lot of toil. Nor would the speaker lose
    anything for having taught you the features -- unless of course
    mushrooms were scarce.
    But in general, when the information you are given through language
    is not about scarce resources for which the speaker and hearer are
    competing, linguistic "theft" is a victimless crime. It's more like
    barter, in which I tell you the things I know and you don't, and
    vice-versa. But it's all about minimising unnecessary toil.
You have clearly though about this issue, though, so you have probably
manage to "steal" enough knowledge about linguistic theft now as you
will ever need!

    (28) Which of the following is an example of procedural memory?
    a. short-term memory
    b. remembering how you found out the answer to this question
    c. remembering when you found out the answer to this question
    d. remembering where you found out the answer to this question
    e. ***none of the above

> I thought that procedural memory is about HOW you remember
> what you do. Why is it not b then? One could argue that I
> tried to retrieve how I learned the difference between
> declarative and procedural memory, not only that they are
> different concepts.

This has actually been queried and answered already:

"This is not a trivial question: Procedural memory is indeed the memory
for how to do something: ride a bike, play a computer game, tie
your shoes, do long division. That's all procedural memory. You remember
how to do it, but you probably don't know how you know how! You've
forgotten the episodes, all those times you tried riding a bike and
failed, and then how you gradually began to be able to do it.
"Now if you understand that clearly, then the word "how" will not
lure you automatically into a reply that something is procedural when
it is not (as above):
"The "How" in "How do you feel?" has nothing to do with procedural
"The "How" in "How did you know that?" is not procedural either (if I
overheard someone say it, that has nothing to do with knowing a
procedure. That's just an episodic memory for a piece of semantic
knowledge ("knowing that").
"It's very important that the concept of Knowing-How vs. Knowing-That
is understood, so that even if the word "how" appears in the sentence,
you understand the difference well enough to see that it is not really
about know-how in the procedural sense at all.
"Now I can tell you how you found out the answer to this question
(if you did!): It was from reading the textbook and lecture notes
and hearing the lecture. Those are all just episodes. You didn't
learn any procedure. You learned some declarative/semantic fact. and
the "how" is just about how you know that fact, not about how you DO
something, which it would have to be if it were procedural memory rather
than declarative/semantic memory.
"The only one who can explain how, from hearing lectures and reading
books, people come to know certain things, is a cognitive modeler who
is trying to explain the mind!"

    (32) Which of the following are short-term memories:
    a. ***episodic memory
    b. echoic memory
    c. iconic memory
    d. digit span
    e. all the above

> What else are echoic and ichonic memory if not short-term
> memory? I thought this is part of Baddeley and Hitch's
> working memory. Echoic memory can be rehearsed whereas
> iconic cannot, and that is why most visual stimuli are
> phonologically encoded. And is it not true that, among
> others, digit span (e.g. 7+-2 chunks according to Miller)is
> used as a way of testing short-term memory?

You're absolutely right (and this question has been junked!).

    (35) The hippocampus is the area where
    a. long-term memories are located
    b. short-term memories are rehearsed
    c. ***long-term memories are consolidated
    d. injury causes retrograde amnesia
    e. none of the above

> I have always thought that information in STM needs to be
> consolidated before getting into LTM. So, is there another
> consolidation process going on in LTM in order to strengthen
> the incoming memory traces to already existing ones?

You're right, and this too has been asked and answered (probably at
about the same moment!) Now that it's come up twice I will mark both c and
e as correct.

"Well again, you clearly knew the full, correct answer, which is that
short term memories are consolidated into long-term ones in the
It's really tough anticipating all the ways a question can be
misconstrued! I had been thinking of c as being like "the dairy is where
the ice cream is made" whereas you were thinking "the dairy is where
milk is made into ice cream!"
"I'll drop the question on your paper, and if more email challenges this
question I'll drop it for everyone."

There were more challenges on the same point, so both c and e are hereby
marked as correct.

This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:53 GMT