Kripke's Meaning Skepticism 3 - A Straight Solution

A straight solution to the paradox is outlined and defended. This solution has been suggested by Craig Delancey, and it says that plus is simpler than quus, and that humans should (and generally do) use simpler rules for producing and recognizing meaningful utterances.

[treyh]

Thank you so much for sharing, you are an excellent teacher and orator.

[dennisfrancisblewett]

@3:08 This seems to be an appeal to numbers (ad populum) argument, whereby the claimant saying that plus is simpler than quus because it's believed it's more abundant in the environment, thus making it a fact of reality that it is indeed simpler. It also appears that thinking with Solomonoff induction might lead to a hasty generalization that is not necessarily true.

[cliffordhodge]

I think the grue/bleen paradigm is more aptly applied in a critique of induction, in which it would likely be put forth: An object is grue if green when observed, blue otherwise, and induction does not let us determine whether something is grue, or MERELY green.   If I say, "X is green before time t and blue after t; therefore it is grue, " this is quite different from the plus/quus problem and from the induction problem.  I can wait until time t passes to see if X is actually grue or "merely" green.  I emphasize "merely" because "X is grue, " and "X is green, " are not competing or incompatible assertions.  I may say, "X is green, " because X is appearing to me green-ly and, "X is grue, " either because after t I am appeared to blue-ly or because I believe I would be so appeared to after t.  In this last case, I can, prior to t, happily assert that X is green and X is grue.  The simple fact that they both have to do with the color (appearance to me) does not make them assertions of the same kinds of facts.  If I have a world view according to which a given object may be different colors at different times, or an object travelling at extreme high velocity changes size or shape, I may use terms for ideas which seem to be more complex than necessary or natural, but which are more simple in that they have a more snug fit with my experience.

[cliffordhodge]

I suspect the plus/quus problem is more apparent than real.  That is to say there really is no problem, but one is offered in a manner that is prejudicial or misleading, as with a Zeno paradox.  There seems to be something along the lines of a hidden dogma of empiricism here.  In discussing whether I can know if I actually mean some proposition A, Wittgenstein considers whether there is something that it feels like to mean A, some identifiable mental state that accompanies an instance of meaning A (being sincere about wanting to inspire in someone a thought corresponding to my statement A, which I believe to be a true statement).  Although he doubts our ability to pin down such a feeling, and despairs of our being able to confidently impute it to others, he does not go so far as to deny that there is such a thing.  This problem is a sort of existential/phenomenal problem very well illustrated in the conversation Dostoevski's protagonist (Notes From The Underground) has in bed with the prostitute, in which he begins to speculate about his possible motives and purposes with this conversation.  One solution to this meaning problem, although not very satisfying, is this: there is something that it is like (or feels like) to mean proposition A, but this something is very basic and abstract, and in fact is made to disappear when searched for.

[absupinhere]

@14:44 And this is how sherlock holmes proves that Kane B lives near a body of water (the background seagull sound)

[cliffordhodge]

Re the coin flip example, it seems this does not illustrate an affinity for simplicity so much as it demonstrates the ability to distinguish between what does and does not bear the mark of mind.  We recognize the first as resulting from the project of a mind, which is important in taking meaning or intension from someone's words.  Given this recognition, one would choose the first string instantly, without any kind of process distinguishing simpler from more complex being necessary, even for purposes of theorizing.  Also, the simplicity test seems reasonable for programming solutions for arithmetic problems, but less so when considering so many other instances of linguistic usage, where sometimes a more ornate phrase or uncommon term conveys an idea more effectively.  If the projects behind the use of language all tend (or are assumed) to be simple rather than complex, then simplicity seems a reasonable measuring stick.

[Ansatz]

Delancey's idea starts with an interesting premise, that meaning could be defined by simplicity, so the simplest interpretation of a word that is compatible with usage is the meaning of that word. The strange thing is that later in the video the idea seems to revert back to Solution 3 from the first video which held that meanings were mental.

12:40 "When we interpret an agent making utterances, we assume that those utterances are caused by some procedures in the agent."

It's rather shocking that the video goes to mental procedures when it is time to justify the choice of simplicity as a definition for meaning. It's like we think that mental procedures are the true source of meaning and we're trying to argue that simplicity closely corresponds to mental procedures.

"If I think that Frank means plus, there must be some procedure going on in Frank that causes the utterances."

The Delancey explanation for meaning is not supposed to work this way. If Frank means plus, the only things we should be able to infer from that is that Frank uses a word in a way that is consistent with plus and where plus is the simplest consistent interpretation. As far as Delancey meaning is concerned, Frank could just as well be a puppet or a bird doing thoughtless mimicry; behavior and simplicity should be all that matter.

"Interpretations of meaning that assign the simplest procedures are likely to be correct, because humans tend to use simpler procedures."

At this point the video is outright admitting that Delancey is just describing a technique for accurately guessing the mental states that form the actual meaning of a word. If Delancey were describing the nature of meaning then there shouldn't be any other thing beyond simplicity that it needs to correspond to in order to be "correct."

We should be arguing that Delancey meaning fits the role that we expect meaning to serve, but apparently we expect meaning to be in the mind of the person making the utterance, as in Solution 3, and so we're forced to show how Delancey's solution closely matches Solution 3. The whole discussion of evolution is an acknowledgement that Solution 3 is a better solution than Delancey's solution because it is apparently so important to show how Delancey's solution corresponds to mental procedures, and so Delancey's solution is only correct to the extent that it matches Solution 3.

23:45 "Simplicity is the mark of the mind."

The mind is Solution 3, not Delancey's solution. It's amazing how quickly the principles of Delancey's solution were abandoned after it was introduced at the start of the video.

"Objection 2: Simplicity is irrelevant. If it turned out that quus was simpler, we wouldn't decide that we meant quus all along."

We _can't_ bite the bullet on this one. Imagine going to a zoo where they have painted mules in the place of zebras. Imagine that you've never seen a real zebra, and whenever you talk about zebras you are actually talking about the painted mules. When you read the sign on the enclosure and it says zebra, you honestly believe that it is refering to the painted mules, and in every way you behave consistently with the painted mules being zebras.

Now what happens when the paint is removed and the mules are revealed for what they are? Just like someone suddenly discovering that quus is somehow actually simpler than plus, you're forced to decide which one you've meant all this time. Does the word zebra for you actually mean painted mule, so the sign was actually accurate for you and every time you spoke of zebras you honestly meant painted mules? If you bite the bullet on this you're effectively saying that you'd continue to use the word zebra to refer to painted mules because that's how you've always used the word and so there's no reason why revealing the trick should change anything.

In reality you have a concept of zebras in your mind and you have a concept of addition in your mind and that is what you mean no matter the situation. Even if you spent your whole life calling painted mules zebras and never once used the word to refer to an actual zebra, you still have an idea of zebras in your head that does not match painted mules. Even if quus were revealed to be simpler than plus by some revolutionary new mathematical discovery, that would not retroactively change the mental process that you'd been using for addition all this time and so it wouldn't change what you'd meant all this time.

Another issue is that Kolmogorov complexity is arbitrary. No reason was presented for why we should use that measure of complexity in place of any other, and even when we're using Kolmogorov complexity we still need to choose the computer language that we're going to use to measure by, and that is also an arbitrary choice.

In the video we choose Turing machines and we choose to represent numbers as strings of 1s, but both those decisions are arbitrary and we're effectively stacking the deck against quus. There are ways to represent numbers which would give quus the advantage and make quus simpler than plus, so our choice of number representation leads to our favored interpretation winning.

[cliffordhodge]

The problem is stated with a cultural bias.  If we imagine, a la' Wittgenstein, a tribe of people for whom there is a word for the counting numbers 1 through 10, but anything over ten is simply "many", or we imagine one for whom the property of being grue is of much greater use and interest than blueness or greenness, so that they call the property of being blue 'grue 1' and being green is 'grue 2', these are both in line with Occam's Razor in that they have fewer, rather than more theoretic entities.

[edvartslogins]

A single coin heads flip is a 50/50 chance, but flipping heads 6 times in a row has less than 50% probability because of all the 50/50s you need to go through beforehand, what am I missing if that's wrong?

[Lucas-bfpw]

Is Kolmogorov metric really objective? I mean, you've made some pretty arbitrary choices on how to represent the numbers for example. Comparing explanations X and Y, what if X is written shortly under one representation, but Y is written shortly under another?

[I_Am_Midnight-i]

Kane, I’ve been watching your videos on modal logic, can you please answer this question? So, we have this argument---> [](P=>Q)=> ([]P=>[]Q). So, we assume its negation and then its derivation and that turns into

[](P=>Q)
~([]P=>[]Q)

My question here is WHY do you ONLY apply the rule for the consequent and not the antecedent? That is why do you only do this---->

[]P
~[]Q

And you don’t include the antecedent, the being

~P
Q

Please answer this!!

[InventiveHarvest]

I think simplification is a rule or procedure. Kripke might say that how do we know what we mean by simpler? We could mean cimpler - the size of a computer program to compress a given string of symbols unless one of those symbols is "0" and then output infinity. A turing machine would additionally have a hard time reducing simplicity on non-terminating programs. How can it determine simplicity between two non terminating programs?

The evolution bit seems subjective to me. i.e we dont like sugar because it is sweet - sugar is sweet because we like it. We cannot see infrared, but dogs can, etc.

Addition is simpler than addition with complex numbers, but I would not assume that someone adding was necessarily throwing out the complex numbers.

Simpler is well defined, but so is "addition"?