Thread: Thoughts on Thoughts on Language

I wanted to start this thread for along time so here goes. This thread has no real central theme (other than cognitive science and linguistics) I just wanted to post some rambling thoughts on language, representation and thought. (Yeah I'm so meta)

Propositions represent a state of affairs in the world which can be broken down into syntactical parts called words. Even these words can be reduced to more fundamental components called morphemes. Morphemes are the most fundamental, semantically meaningful units in words and are atomic and irreducible. All words are composed of one or more morphemes.

Note that morphemes are distinct from syllables because the latter refers to speech sounds not syntax and semantics like morphemes.

For example the word "tallest" is composed of two morphemes, "tall" and "est". "Tall" is a free morpheme because it is (semantically) irreducible yet it can convey meaning on its own. "Est" is a bound morpheme because it is semantically meaningless unless it is accompanied by a free morpheme.

What I am interested in is how free morphemes express meaning. (Or a more broad question of interest to linguistics and philosophy, is how words express meaning) We can ascertain the meaning of a proposition by reducing it into syntactical parts but morphemes cannot be broken down in this way so how do they express meaning?

The word "tall" does not refer to an object, it is an abstract conceptualization of objects. Neither does it refer to one thing or a state of affairs because the notion of height is relative to the frame of reference of the observer. How is the complexity of a concept like "tall" symbolized by such simple arrangements of words (which are seemingly arbitrary)?

When I say the word "tall" (even though it is complex and has semantic baggage as I suggested above) you instantaneously understand what I mean. How is it that the human mind can comprehend such abstract concepts (like tall or infinity or nothingness) through symbolic words? Is there an innate faculty that enables us to deal with abstraction that is absent from other animals?
What is the nature of the representation between language and thought? Does language act as a mirror to reality or is it arbitrary? <Same question applies to thought.

Feel free to answer some of my questions or add your own.

I really don't know very much about the current state of this field... I have opinions but I could be being very naive.

I am sceptical that language is what structures thought (and isn't that a common claim?), although it's not entirely clear to me what 'language' is defined as. But in a superficial sense of the word, I would say that language is probably a layer on top of what makes humans special (let's call it abstract thought); language simply facilitates the expression of that thought.

I think human thought is fundamentally pattern recognition and manipulation. We learn the word 'tall' by seeing a number of objects with varying properties but the shared property of tallness, and we keep hearing our parent make the sound 'tall', and then we associate the pattern with the sound.

I therefore share the analysts' position that concepts like nothingness are out of the domain of human knowledge, and talking about nothingness and its properties is literally meaningless; the overestimation of the realm of human thought is behind most problems of metaphysics. I am not sure if the analysts share my view of what human thought actually is, though.

I was literally thinking about infinity in this context only yesterday. I think it is something of an anomaly. It's quite a nuanced and technical issue though; in the late 19th century mathematicians pinned down infinity in terms of finite concepts.

Thank you for the response Xei. I don't really know much about linguistics either, which is why I made the thread. It is indeed a common claim that language structures thought and while I do think this is a valid assertion (I think many people agree that there is a correlation between the anomaly of language among homo sapiens and our immense cognitive abilities), I am also interested in the ways in which thought structures language or how our perceptions structure language.

This is primarily the reason why I made the Kiki/Bouba thread. I want to understand what it is about the world that causes us to talk about it in a certain way. Wittgenstein (in the Tractatus but later refuted in the Investigations) believed that language acted as a mirror that reflected the logical structure of the world. I have my own criticisms of this but I don't really feel like getting too deep into Wittgenstein in this thread so I will just say this; he asserted a logical ontology/metaphysics, in that the "world" can be broken down into states of affairs or logical relationships like "the cat is on the mat" which I suppose could be represented by E! cRm (the E being an existential quantifier to say there exists one cat in relation to the mat). The logical nature of language acts as a mirror to reflect the logical nature of the world. I don't exactly agree with this position which I will try to explain later. Back to Kiki/Bouba.

The (perhaps at best dubious) implications of the KB effect suggest that associating speech sounds with abstract shapes or concepts is not arbitrary as we have once believed but seems to be (possibly) an innate cognitive structure present in humans regardless of sex, age, culture. I say that because some 90-95% of people (irrespective of age, sex, culture, native language, etc) correlate Kiki with the jagged shape and Bouba with the curved one. Autistic people statistically showed no preference apparently. To me the fact that an overwhelming majority of people correlate the words with the shapes in a particular way suggests to me the existence of a universal grammar. If speech sounds are not arbitrary in relation to what they represent signifies a first language (which seems pretty intuitive anyway, Ill call it Babelese!) from which all subsequent languages stem from. I think investigation into this theoretical language can give us insight into many problems in linguistics today as well as providing a window into the cognitive structures of the human mind.

I completely agree with your assertion that language acquisition is primarily pattern recognition, Ive also heard you in different contexts say that consciousness follows the same operation which is extremely intriguing to me. This reminded me of these monkeys I have been reading about which apparently have the ability to communicate with "symbolic language" not only with each other but with other monkey species as well. I'm not talking about apes, I mean old world monkey like Diana monkeys and Vervet monkey. Apparently they have different pitched calls for, for example a tiger, all of the monkey upon hearing this call jump up into the trees. When they make the calls for an eagle, they all jump down from the trees onto the ground. To me this implies that they do indeed have a language (albeit a primitive/pragmatic one) similar to homo sapiens. Of course this is contested in linguistics. Maybe the monkeys are conditioned through experience to correlate certain pitched sounds with entities like tigers or snakes, etc (or more generally “danger!”) but is not this the way in which humans learn language?

I kind of went on a tangent; I came into the thread trying to talk about Wittgenstein and theories of meaning. Anyway I wanted the thread to have no real direction so I suppose this post will suffice.

Originally Posted by Xei

I therefore share the analysts' position that concepts like nothingness are out of the domain of human knowledge, and talking about nothingness and its properties is literally meaningless; the overestimation of the realm of human thought is behind most problems of metaphysics. I am not sure if the analysts share my view of what human thought actually is, though.

I was literally thinking about infinity in this context only yesterday. I think it is something of an anomaly. It's quite a nuanced and technical issue though; in the late 19th century mathematicians pinned down infinity in terms of finite concepts.

I would tend to agree with you that talking about nothingness is meaningless, in other words we can talk about it all day and never come to a consensus because we only experience, well somethings, nothing is an abstraction. I guess I should not beat that dead horse anymore than I already have.

I think about it like this: Imagine two people are asked to figure out the contents of a closed box without opening it. Person A says it is a pencil; Person B claims it is a bunny rabbit. But this dispute is trivial, they can argue all day about the contents of the box without coming to a consensus but unless we can actually empirically verify what is in the box, anyone’s guess is as good as anyone else’s. I could say it is the Flying Spaghetti Monster (bless his noodles) and no one would be able to prove me wrong without empirically falsifying my assertion. Science and empirical verification give us the tools to open the box and peer into its contents.

I think this is articulated more clearly by a greater mind than mine: William James. This is an excerpt from his essay “What Pragmatism Means” which I think really exemplifies the pragmatist position regarding metaphysics.

“What difference would it practically make to anyone if this notion rather than that notion were true? If no practical difference whatever can be traced, then the alternatives mean practically the same thing, and all dispute is idle. Whenever a dispute is serious, we ought to be able to show some practical difference that must follow from one side or the others being right.”

Wittgenstein held a somewhat similar position regarding metaphysics and meaning. In the Tractatus he asserts (as I alluded to in my previous post) the meaning of a proposition can only be determined through empirical verification. While I don’t necessarily agree with this correspondence theory of truth I am sympathetic towards it. This is why he rejected metaphysics and ethics on the grounds of them being meaningless. Take the metaphysical proposition “God exists” and an ethical proposition “We ought not to harm others”, neither of these propositions corresponds to any state of affairs in the world and we can only meaningfully talk about what we can experience.

I would be way out of my league in a discussion about the concept of infinity with you so I will try not to stick my foot in my mouth too much. I believe infinity to be an abstraction from experience. For instance all natural numbers can be squared and those squares can be squared as well, ad infinitum. Because of this property of numbers we can conclude that “infinity exists” (whatever that means). The word infinity does not correspond to an object because it is constantly in flux by definition. I don’t completely rule out the possibility of infinity but if it does “exist” then I don’t think we could talk about it meaningfully anyway. Words are static (but yes meanings sometimes change over time) and I am skeptical of the notion that finite words can symbolize something that is so abstract that we can barely conceive of it.

I have been thinking about the common analogy between language and DNA which is where some of this loosely stems from. There is an extraordinary diversity among biological life due to the complex diversity of specific arrangements of only four nucleotides. Simple arrangements of these four fundamental nucleotides are what distinguish a tree from a squirrel; a bee from a flower. There are millions of books composed of arrangements of only 26 letters while there are millions of species of animals and plants composed of only 4 nucleotides. We can use reductionism to understand biological systems in parts but can we use the same methods to understand the nature of language?

Can we answer the question of meaning through breaking down propositions into semantically smaller parts until we reach the bedrock of language (morphemes)? A recent shift in biology has in part turned away from reductionism to systems biology which investigates the ecosystem or organism as a whole. I believe both methods are useful in there own right but what about language?

Does the free morpheme “tall” (which is atomic and irreducible) convey meaning without being accompanied by other words in the language like “small”? Let me phrase that a bit better. I suppose I am postulating a kind of linguistic holism. Words have meaning only by virtue of being accompanied by other meaningful words in a language. For example the word “Father” is only meaningful if it is accompanied by words like “mother, son, and daughter”. In a way you could say a word is defined by what it is not, almost paradoxically. This is the first most obvious criticism of this position that comes to mind.

Second, we have the problem of foundationalism. If words meanings “build off of each other” then how is it that the “first word” conveys meaning?

In conclusion I think we have to look at language as a whole and analyze how the parts interrelate to create the emergent property of meaning.

I'll try to be concise. Please don't mistake it for brutality, everything has been carefully considered.

Originally Posted by stormcrow

The (perhaps at best dubious) implications of the KB effect suggest that associating speech sounds with abstract shapes or concepts is not arbitrary as we have once believed but seems to be (possibly) an innate cognitive structure present in humans regardless of sex, age, culture. I say that because some 90-95% of people (irrespective of age, sex, culture, native language, etc) correlate Kiki with the jagged shape and Bouba with the curved one. Autistic people statistically showed no preference apparently. To me the fact that an overwhelming majority of people correlate the words with the shapes in a particular way suggests to me the existence of a universal grammar. If speech sounds are not arbitrary in relation to what they represent signifies a first language (which seems pretty intuitive anyway, Ill call it Babelese!) from which all subsequent languages stem from. I think investigation into this theoretical language can give us insight into many problems in linguistics today as well as providing a window into the cognitive structures of the human mind.

I would tend to disagree very much with this interpretation. I would be hesitant to extend the Kiki/Bouba thing anywhere beyond what it simply is, which I regard thus: for aesthetic reasons, humans like to associate things with similar things (through their pattern recognition faculties). In the case of kiki, sharp changes are inherent in the physical sound, and hence there exists an objective similarity between to sound 'kiki' and the kiki object. Giving units of language a sound similar in some sense to the object is just done for practical reasons, and because it amuses us. In fact it's exactly analogous to the letter k itself, which is sharp, as opposed to b and o, which are rounded. This helps learners remember what they mean, and is aesthetically pleasing. Obviously there is no 'universal alphabet'; it is simply that the visual similarities form a universal substrate (as opposed to an internal thing).

I'm not exactly clear what 'universal grammar' means. In the crude sense above, it is patently obvious to anybody who has studied a language that grammar is not universal. In other languages, adjectives can precede or succeed nouns, verbs can follow pronouns or be sent right to the end of the sentence dependent on tense.

In a more general sense of a universal logic, again I would say that it is extrinsic. We all perceive the same world and hence the same patterns in it (the very general and important ones, anyway), and so we all generate the same mental models. Implication is a classic example. There's not really any such thing, it's just a pattern in experience.

Monkeys

I don't think there's much to be said here... other animals certainly have language, although none of them come remotely close to human language. You can actually do statistical analysis on the sounds animals make to put an objective figure on the complexity. I really can't remember clearly but I think dolphins are about halfway to humans, but this is on a logarithmic scale. Primates have only ever been observed to use single morphemes, which represent obvious patterns (such as 'danger' or 'food'). Undoubtedly what they are doing is a prototype of human language. In fact it could even be the whole of language, if language, as I speculated, is just a representational layer on top of conceptualisation; it's just that their thoughts are primitive. It is worth bearing in mind that humans are miles ahead at both conceptualisation and language. It seems unlikely that we evolved two unique traits so quickly, so one would imagine that there is a causal link.

Originally Posted by stormcrow

What difference would it practically make to anyone if this notion rather than that notion were true?

I have heard something very similar from Ayer, I think, and Hume basically says it (personally I think he 'basically said' all of the philosophical truths which followed him).

I'm not sure if it is equivalent to my beliefs, but it seems to fit. There are three classic metaphysical questions I like to use as examples as things which are actually meaningless when you consider the pattern recognition basis of thought, and asking the above question of them places them in the same category:

Is there an objective world?
Is there free will?
Why is there something rather than nothing?

When I analyse the meaning of these words and sentences, I find that they are meaningless, and it is true that you cannot possibly observe anything that would bear upon them, so by that criterion they are also meaningless.

I would be way out of my league in a discussion about the concept of infinity with you so I will try not to stick my foot in my mouth too much. I believe infinity to be an abstraction from experience.

I think you're right, basically what mathematicians (Weierstrass was the main one) did was formulate infinity in finitary terms. 'Space is infinite' simply means, given any point in space, you can walk a meter in any direction from that point. It's not actually a concept about the totality; it's about finite, local concepts. 'The function tends to infinity at a' just means, if I ask you for any ridiculously huge (but finite) number, you can make the function greater than that number in some positive region around a. This can be proved, and the idea of infinite quantities is involved nowhere. It is of crucial importance to understand that the function is not actually defined at all on a. To get a feel for it I would suggest you try to prove that the function f(x) = 1/x^2 (defined for every x except 0) tends to infinity at 0.

Originally Posted by stormcrow

Can we answer the question of meaning through breaking down propositions into semantically smaller parts until we reach the bedrock of language (morphemes)?

This is what I was trying to explain in that conversation with really; unfortunately he just got angry and gave up before I explained.

Again, I take the opposite view. I think that reductionism works beautifully in my philosophy (which again is just Hume from a modern perspective). There is no need for holism: the word 'tall' exists independent of the word 'short', and the foundationalist problem is not, for me, a problem. Not in the realm of language, anyway.

The way it works is simply by pattern recognition. The concepts in your brain are just patterns, symbols for commonalities in experience. When you really appreciate this idea, I think it's as beautiful as when you really appreciate the idea of natural selection. It makes so much sense from a naturalistic perspective... there is no magic involved in human thought, it's just a simple algorithmic process, and one which is really a very obvious adaptation.

Talking about the 'meaning' of atomic words (I agree totally with your analysis from compound words down to atomic ones, I just embed the atomic ones in the rock of experience) is a bit of a misnomer: a word just is the pattern it designates. We see many scenes with a variety of shapes: some of these shapes are bigger than ourselves. Thus we acquire the word 'tall'. The word 'tall' doesn't exist by virtue of the word 'short', it simply designates any object which satisfies the pattern. The word 'red'; there's a whole continuum of colours, but some of them always activate a specific bunch of neurons; thus we acquire the concept for red. If somebody is colour blind to all shades but red, they still acquire the concept of red (again, bearing in mind what exactly what the 'meaning' of 'red' means).

Oh. That wasn't concise at all.