Why is Jason so utterly stupid? The answer is that he studied and then taught stupid shit. In particular, nobody told him that arbitrary assertions, though useful in certain contexts, have no 'naturality'- i.e. there are no 'natural transformations' such that a particular map between functors can be done consistently over an entire category. Thus, though useful, arbitrary assertions don't enable you to make any general or 'nomothetic' point.
Stanley published a paper titled 'Making it Articulated'- which sounds ungrammatical but is actually meaningless unless you think chopping up and then reassembling a thing 'makes' rather than destroys it- more than 20 years ago. Yet, even at that time, it was obvious that 'Transformational Grammar' had failed. There is no 'natural' or non-arbitrary way to understand a particular language or to translate it perfectly into another. That is why, though the Bible is the most precious book in the English language, there are many different translations of it- all done by very learned and pious people.
The big problem with trying to philosophize about language is that we don't know all the possible meanings of any utterance. We arbitrarily jump to a particular conclusion but somebody else could easily correct us- if it were worth their while.
Consider the following-
(13) Everyone, who John ran, he liked.
Clearly John was running agents for an intelligence agency. He liked all his agents. Indeed, this was one reason they liked him and did his bidding.
Stanley assumes this is an ungrammatical sentence. It may be but it may not be. One may make an arbitrary assumption for some useful purpose but nothing general can be deduced from this.
If there were the sort of pragmatic process that allowed an interpreter to smoothly replace an uttered sentence by another sentence that contains a new bound variable, the ungrammaticality of (13) would be a complete mystery.
The sentence isn't ungrammatical at all if John was a Le Carre type spy. Jason insists otherwise because he is not aware that, at MI6, a spy runs agents.
What about- ' Everyone has had the privilege of having John greet'?. Is it 'wildly ungrammatical'? It doesn't sound right. We were expecting to hear 'greet them'. But it is possible that John's method of greeting is highly entertaining to watch. Having him greet is indeed a rare privilege, though, being a simpleton, John tends to greet only items of furniture. Sadly his method of bidding adieu is highly insanitary and not safe for public viewing.
Can Jason provide us with any logical arguments as opposed to absurd, ignorant and utterly arbitrary stipulations? Let us see-
Abstract:
I argue in favor of the view that all the constituents of the propositions hearers would intuitively believe to be expressed by utterances are the result of assigning values to the elements of the sentence uttered, and combining them in accord with its structure.
We hear other people saying things all the time. We never 'assign values' to the elements of a sentence unless we are required to do so as part of our job or else are a not a native speaker of the language in question. We don't understand some sentences we hear and we ignore other sentences. We may 'get the gist' for some purpose of our own but we may, equally, deliberately misunderstand things or 'hear what we want to hear'.
The way I accomplish this is by questioning the existence of some of the processes that theorists have claimed underlie the provision of constituents to the propositions recovered by hearers in linguistic interpretation, processes that apparently bypass assigning these constituents to elements of the logical form of the expression uttered.
Why bother? Those theorists were useless. Even twenty years ago, it was clear that progress in 'speech recognition', automatic translation, etc would be based on big databases and sophisticated Statistical techniques. 'Greedy' algorithms are too costly, heuristic algorithms or 'self-learning' black boxes perform better. The Philosophy of Language turned out to be as utterly useless as the Philosophy of Mind. We can make a machine which is articulate enough for a specific purpose. But we may not be able to say how it is 'articulated'- i.e. put together. There may be no way to discover what happens within a 'black box'. All we can say is whether it works and is useful or whether it does no such thing.
Here is the view of linguistic communication I find plausible.
This was written 20 years ago. At the time few would have thought this 'Philosopher' would think it plausible that everybody he didn't like was a fucking Nazi.
First, a speaker makes an utterance, and her linguistic intentions uniquely determine a certain syntactic structure, or ‘logical form’, as it is known in syntax.
What are 'linguistic intentions'? How are they different from 'intentions'? It may be that there are certain avant garde poets who have 'linguistic intentions' unrelated to anything they actually want or desire. But even in that case what 'uniquely determines syntactic structure' is the knowledge base concerning syntax and what type of verbal behaviour is appropriate in a particular situation. This why, during my brief career as an internet Guru, I sounded like Yoda. Then, when I found my Honeytits Cumbucket persona was getting more hits, I switched to Valley Girl vocal fry.
If her utterance is a successful linguistic assertion, the logical form is sentential.
So 'me love you long time' is sentential. Who'd have thunk it?
Successful interpretation involves
understanding what the other guy is getting at. You don't need to know the guy's language. Having a good 'theory of mind' and being able to intuit what the guy needs or desires is what is important. When I say 'where is the little boy's room' I am not in fact asking the way to your child's bedroom. I want to use the toilet. But you can guess that by the way I am hopping from one foot to the other.
assigning denotations to the constituents of the logical form, and combining them in accord with composition rules that do not vary with extra-linguistic context.
Fuck off! The only thing that matters is the 'extra linguistic context'. Language is merely a tool.
The denotations that successful interpreters will assign to constituents of a logical form will be constrained by the linguistic conventions governing those elements.
No. Some phrases or collocations become 'conventional'. But we each have some unique expressions in our idiolect which we think are cutesy or demonstrate 'authenticity'.
We successfully interpret Yoda's fractured syntax. Indeed, we could even understand Dubya's mangled English. What we don't understand is the fine print in contracts drawn up Ivy League educated lawyers working for White Shoe firms.
In the case of certain elements, which wear their context-dependent nature on their sleeve,
all elements of any utterance do so unless the thing is gibberish.
the linguistic conventions governing them are rather lax.
Anything at all is a 'convention'. The thing is wholly arbitrary. At one time, it was conventional for analtickle Philosophers to keep abreast of developments in Science and Technology. Then they just created a citation cartel for the terminally stooopid.
These elements are the obvious indexicals, the obvious demonstratives, pronouns, context-dependent quantifers such as many and (perhaps) that, and covert pronominal elements whose existence can be demonstrated by purely syntactic tests.
The guy is now pretending that he knows from grammar and linguistics. But then Chomsky tried to turn Linguistics itself into gibberish wholly unconcerned with any actual language.
What results from a successful application of this first stage of interpretation is a unique proposition, a fully truth-evaluable entity.
Nonsense! There may be a most probable interpretation for some specific purpose but it is not unique. Suppose Jason and his barmy army really had found a way to get unique interpretations, then they would be expert witnesses in important cases involving Constitutional law. Can you imagine Jason testifying in the dock re. Roe v Wade. First he would denounce the Judge as a Nazi before explaining that Science is just Fascist propaganda. There is no such thing as a foetus. You can't kill what doesn't exist.
Furthermore, if the process is successful, the proposition it yields is one the interpreter would recognize as the proposition expressed by that assertion.
So, the rapist is justified in interpreting 'No! Get the fuck off me!' as 'Yes please! I'm totes into this Madam Merkel'.
In the final stage of a successful communicative interchange, an interpreter applies standard Gricean reasoning to derive the implicatures of the linguistic act.
Fuck off! Gricean implicature is 'anything goes'. Any way, some people will be better at grasping intention even if they don't know the relevant language. Thus a lady who is pretending to find me witty or well informed so as to convert me to her cult will be immediately understood by any other woman present as doing a bad job disguising her loathing for big fat bore that I am. The smarter ones will also understand that I am trying to recruit the girl to my Socioproctological Institute. When I say 'mine', obviously what I mean is that I founded it but have currently been suspended from membership because of allegations of sexual self-abuse. Currently the only member, that too part-time is the neighbour's cat which sometimes jumps through the window. Sadly, it doesn't believe me when I protest my innocence of the various charges I keep bringing against myself.
Virtually every claim I have made in the description of my project has been severely criticized in recent years, most prominently by Kent Bach, Robyn Carston, Francois Recanati, Dan Sperber, and Deirdre Wilson.
These really aren't smart people. They haven't made millions consulting for Alexa or ChatGPT.
In particular, these theorists hold that in many cases of successful interpretation, the prop-osition that the hearer would intuitively believe to be the proposition expressed by the communicative act contains elements that are not the value of any constituent in the sentence uttered, nor introduced by composing these values. Instead, these elements are provided directly by context.
No. They are provided by things internal to the interpreter including their guess as to what is the context. Thus, properly understood, my allegations of sexual self-abuse were not simply a means to drum up a bit of publicity for the Institute of Socioproctology at a time when the 'Me-Too' movement was at its height. The truth is, I was hoping for a big pay-off from the Institute. Sadly, the neighbour's cat hasn't yet paid its membership fee and so I have not yet been able to retire to the Costa del Sol.
Loosely following John Perry, I shall call such elements unarticulated constituents
In which case everything that has ever happened in the Multiverse as well as everything which couldn't possible happen are elements of every utterance. What is the point of having an 'intension' whose extension is everything conceivable or inconceivable?
If there are unarticulated constituents of utterances,
Then I uttered the King's speech in Parliament. Also, my saying 'Mummy, gimme bis-cute' back in 1967, actually contained all the great discoveries contained in STEM subject Nobel Prize winners' acceptance speeches.
then the result of combining the values assigned to elements of a structured representation is at best an interpretive way-station on the path to the truth-conditions hearers would intuitively associate with the linguistic act. One consequence of relegating the result of semantic interpretation to this role is that intuitions speakers have about what is expressed by linguistic acts are no longer an accurate guide to syntactic structure and semantic content (cf. Bach, forthcoming)). In contrast, I believe such intuitions are an accurate guide to syntactic structure and semantic content (cf. Stanley and Szabo, 2000a, p. 240).
Nothing wrong with having a belief which is useful to you in your line of work. If 'syntactic structures' is what you study, you may as well believe there is an intuition which can grasp it. Otherwise, how would anybody have any knowledge about it? This is convenient but arbitrary. But then, au fond, everything is arbitrary and arises from 'uncorrelated asymmetries'.
But I recognize that this position needs to be argued for, and not simply assumed.
Nothing wrong with making a simplifying or otherwise useful assumption. But it is pointless to argue that the thing has 'naturality'- i.e. is non-arbitrary.
My purpose in this paper is therefore to raise doubts about the existence of unarticulated constituents.
Why bother? The thing is silly. It is obvious that people mean more than they say though what they say, most of the time, is as meaningless as the cat's miaow. My point is that all members of the Institute of Socioproctology- except those who have been suspended- should pay their dues regularly- preferably by direct debit. Saying miaow simply won't cut it no matter what its 'unarticulated constituents' might be.
That is, I want to argue in favor of the view that all the constituents of the propositions hearers would intuitively believe to be expressed by utterances are the result of assigning values to the elements of the sentence uttered, and combining them in accord with its structure.
In which case the hearer's 'intuition' is doing the heavy lifting. But intuition is not something we can lay our hands on or describe analytically. How do we know if an intuition involves just the evidence that is presented and not everything else in the knowledge base? Is inutition some divine faculty free of any personal prejudice or bias we might have?
The way I shall accomplish this is by questioning the existence of some of the processes that theorists have claimed underlie the provision of unarticulated constituents to the propositions recovered by hearers in linguistic interpretation.
It is easy to knock down 'processes' suggested by useless cretins. Why not look at 'processes' used by language recognition and machine translation computer programs? If you are smart enough to improve such programs, you can get very rich. You can still do philosophy- because there are obvious 'open problems' in the underlying math- but people will respect you because they will know you are Elon Musk level smart. You may still be crazy and see Nazis under the bed, but people will give you a respectful hearing and separate out the paranoia from the sound reasoning.
In the first section, I introduce an argument I have exploited in recent work against the existence of unarticulated constituents.
It is nonsense. Jason arbitrarily assigns to some element in the utterance the same value as the supposed unarticulated constituent. Suppose I deny the existence of cats. A cat jumps through the window and comes and sits on my lap. It says 'miaow'. I explain that the 'mi' constituent of 'miaow' causes birds to take on the appearance of cats, even retrospectively, when they fly through the window. Thus no cats actually exist. If they did they would pay their fucking membership dues already.
In the second section, I discuss an influential paper of Wilfred Sellars, which has led some philosophers of language to the view that there is a legitimate notion of ellipsis that is the process by which unarticulated constituents can be provided. If there is such a process, then it can be exploited to respond to the kind of argument raised in the first section. However, I argue that no evidence has been given by Sellars for the existence of such a process.
But ellipsis occurs all the time! We know this. The evidence pre-exists. Interestingly, the Sanskrit word for it is the same as the word for 'dispersion'. The notion is that our attention is pointed to something not in the text before the text itself resumes.
In the third section, I sketch some strategies theorists have appealed to in defending the existence of unarticulated constituents. Finally, in the fourth section, I argue that the processes that one needs to postulate to evade the arguments against unarticulated constituents described in section one lead to unacceptable consequences.
It is one thing to describe and explain features we find in language- more particularly foreign or ancient languages. This can be useful. It is another thing to pretend that those descriptions or explanations have a concrete existence. An arbitrary stipulation may be useful. But arbitrary stipulations against arbitrary stipulations are useless.
1. The Argument from Binding
In recent work, I have given the following sort of argument against the exist-ence of unarticulated constituents. First, I consider a construction utterances of which allegedly involve unarticulated constituents. I then show that there are readings
arbitrary readings. I can interpret 'miaow' as a promise to pay membership fees to the Institute of Socioproctology.
of the construction in question that require recognizing that a variable exists in the syntactic structure of the relevant construction,
like the 'mi' in 'miaow'
whose value, relative to a context, is the allegedly unarticulated constituent.
which is why cats are actually birds.
If so, then the constituent is not unarticulated after all.
Which is why when I said 'Mummy gimee a bis-cute' I was actually outlining David Deutsch's 'constructor theory'. I'm not really a moron. It's just that nobody knows how to properly evaluate the constituents of my speech.
Here is an example of the argument at issue. An utterance of (1) can communicate the proposition that every bottle that John just bought is in the fridge:
(1) Every bottle is in the fridge.
No. John may have drunk and discarded a bottle or given one to the neighbour.
That is, an utterance of (1) can communicate a proposition about a restricted domain of bottles.
It can communicate anything you like though others might think you are being stoooopid.
For example, an utterance of (1) can communicate the proposition that every bottle Sally bought on April 11, 2001 is in the fridge.
No. It is obvious that Sally may have broken one or given one away. We may be entitled to assume that all the bottles bought by Sally which are currently in the house are also in the fridge.
Let us call this phenomenon quantifier domain restriction.
Why? The fact is, a domain is only a domain if it is well defined. 'Bottles' is not well defined. Is a 'ship in a bottle' a bottle or a work of art which does not belong in the fridge. Sometimes, when we use the word bottle we also mean cartons and cans. I may say 'let's open a nice bottle of Chardonnay' though I very well know that any wine we can afford to buy comes in a fucking box.
There is no harm in using mathematical terminology provided you follow mathematical protocols. Pretending that a particular 'domain' is well-defined when it isn't will quickly lead to fallacious reasoning.
On one natural account of the phenomenon of quantifier domain restriction, the syntactic structure of (1) does not contain a variable whose value,relative to a context, is a restricted domain.
That is highly unnatural. Fridges are known to have limited capacity.
Rather, denotation assignment and composition yields the proposition that every bottle (unrestricted) is in the fridge. Successful interpretation then involves augmenting this minimal proposition with material from extra-linguistic context.
'Fridge' is linguistic. We know fridges are small relative to the number of bottles in the world. Why is Jason pretending otherwise.
Call this the prag-matic account.
What other sort is accessible to us. Language is useful and though it can be useful to linguists, or those studying a foreign language or those who wish to write in a particular style, to have an grammatical analysis of language, sill that itself is just a matter of pragmatics.
There may have been a time when some believed that a person who knew grammar sufficiently well would be able to cast spells and work magic. But that time has long gone.
The pragmatic account of quantifier domain restriction can be challenged with the following sort of argument, which I shall call the argument from binding.
Which is like an argument from flying unicorns or other things which don't exist.
Consider a sentence such as:
(2) Every student answered every question.
An utterance of this sentence may communicate the proposition:
(3) Every student answered every question on her exam
Not if we are speaking of a progressive school where there are no exams. The questions students answer have to do with whether they went potty or whether they want to go potty.
These facts
What facts? (3) does not follow from (2)
challenge the pragmatic account; and this in two related ways. First, these facts provide evidence for a covert pronominal element, the content of which, in this context, is the same as that of the italicized material in (3).For positing such an element,
an arbitrary and unwarranted procedure. Suppose I am employed as a teacher. The Headmaster would ask me whether I had actually set any exam and, if so, whether I had actually gone through the scripts handed in by the students. Even that would not be enough. The headmaster would insist on knowing whether the questions I had put in the paper could be answered by drawing obscene pictures.
and holding it to be bound by the quantifier every student delivers the relevant reading (more on this below). Secondly, these facts challenge the pragmatic account, because it is not clear how to capture the bound readings of such sentences given the resources available to the advocate of the pragmatic account.
The 'bound reading' is just an arbitrary and unwarranted assumption.
Let me explain each of these points in turn. Much syntactic structure is unpronounced, but no less real for being unpronounced.
No. Syntactic structures are hypotheses. They are not real 'things in the world'.
This fact raises the question of how to detect unpronounced syntactic structure.
Invent it and then claim to detect it. This how Jason is able to prove everybody he doesn't like is a Nazi.
A working hypothesis is that an unpronounced element exists in the structure of a sentence
but that 'structure' is itself a hypothesis- a useful one for people learning foreign languages.
just in case there is behavior that would be easily explicable on the assumption that it is there, and difficult to explain otherwise. For example, here is a classical argument that the agent of a passive is actually still syntactically present, though unpronounced. In control theory,
i.e. Transformational Grammar which turned out to be stupid and useless and thus was abandoned. Chomsky is a paranoid nutjob just like Jason. Given enough time, they will denounce each other as Nazis.
the theory of the nature of unpronounced elements like PRO in English (the subject of infinitival clauses), there is strong evidence that ‘PRO must have a syntactically local controller.
But this does not generalize across languages. It was just a case of pretending to find evidence you had yourself planted so as to appear to be smart.
This, after all, is what explains the ungrammaticality of:(4) *The ship sank to collect the insurance.(5) *The record broke in winning the race.(4) and (5) are ungrammatical, because
they don't sound right to native speakers. Otherwise they are perfectly fine. After all, we may subscribe to a theory of pan-psychism whereby the ship had the intention to sink in the belief that it would be able to collect the insurance money.
their syntactic structure is:(6) The ship sank [PRO to collect the insurance](7) The record broke in [PRO winning the race]
Records may be said to compete with each other for the prize of being the bestest record ever. My record for loudest farts frequently competes with my record for smelliest fart. Then the former also broke the latter. Sadly, I had shat myself and thus it was disqualified.
And these occurrences of PRO
they don't occur. They have been arbitrarily inserted.
have no potential local controller; or rather, the available local controllers (‘
the 'ship’ and the 'record') are not expressions that denote things capable of collecting insurance or winning races.
why not? When speaking we are welcome to use any manner of speaking. Suppose the ship is operated by an AI. It may calculate that sinking itself and collecting the insurance is what is in its best interests. The Guinness Book of records may decide to hold a 'race' such that the record which gets most hits over a specified period wins a prize. The guys who manage the 'smelliest fart' department are competing with the 'loudest fart' department. But for my shitting myself and thus getting disqualified, my record for loudest fart would also have broken the smelliest fart record.
However,(8) and (9) are perfectly in order:(8) The ship was sunk to collect the insurance.
but the days of sentient AI operated ships aren't too distant.
(9) The record was broken in winning the race.
That sounds odd. The winner of the race also broke the current world record.
This strongly suggests that the agent by-phrase in a passive is in fact syntactically present after all.
Stuff that strongly suggests itself to a cretin is cretinous shite. The plain fact is that an arbitrary insertion of words into a sentence can be useful. But that is all it can be. 'Naturality' is off the table.
This is a standard way to argue that a constituent that is not phonologically realized is nevertheless syntactically present.
For some particular purpose it might be useful to deem that this is the case. But the thing is arbitrary.
The argument from binding is an argument of the very same structure.
It is arbitrary but also utterly useless and easy to contradict. I'm not exactly a genius and I've just kicked Jason in the slats repeatedly. On the other hand, it must be said, Jason probably didn't put any mental effort into writing his papers. Why bother? Only morons go in for this shit. Anyway, his real passion is finding Nazis underneath his bed.
One characteristic syntactic feature of pronouns is their capacity to be bound by variable-binding operators.
Everything can be said to be bound by such operators. But saying stuff don't make it so.
By demonstrating the existence of bound readings of quantifier-domain variables,
One can't do any such thing. All you can do is come up with a better computer language or a better machine translation system of something of that sort. You will earn mega-bucks. You won't have to teach stupid shit to thickos.
one provides evidence of behavior that is explicable on the assumption that there is an unpronounced pronominal element, and difficult to explain otherwise.
The explanations given by shitheads like Jason are of no interest to anybody.
This raises the second challenge to the pragmatic account, which is to explain the bound readings of the sentences in question without postulating a variable in the syntactic structure of quantified noun phrases.
This is like challenging a pragmatic account of Jason's stupidity by postulating an invisible flying unicorn which farts out invisible rainbows which form themselves into the words 'Jason isn't stoopid. He's actually a very clever little boy' above Jason's head.
No comments:
Post a Comment