Do we think in words / language?

[Fruitloop]

Another point is that language inside the brain is necessarily different to language as expressed in sound or writing. Spoken language has to be linear, there has to be an order of words (interestingly, sign language doesn't have to be linear in this way), mental language isn't constrained in this way.

Early behaviourist psychologists discovered that the mental voice we speak to ourselves in actually produces very tiny movements in the vocal tract, leading them to claim that our 'mental voice' was in fact just us talking to ourselves very quietly and that all thought was in fact a result of language. This was only refuted by a psychologist who deliberately paralysed himself with curare to prove that he was still capable of thinking while his vocal tract was incapable of movement (he had breathing apparatus of course).

The brain isn't a computer in any way resembling the machine that I'm writing this on. But neither is the brain magical, it is a physical organ. Language must have a representation at the mental level which is different from its appearance at the oral/aural level. The exact nature of that representation is still a matter for research.

He would have needed to paralyse his Broca's Area as well I think.

 

[weltweit]

I think other chimps besides us rely heavily on gesture to communicate, vocalisations are mostly used for emotive expression or warnings - we are unusual in relying on them for the bulk of our communication. I reckon this is why people can take to signing quite easily, and also that there are vestigial elements of gestural communication in dance, music and architecture.

But there is a body of thought that humans do still rely strongly on non verbal communications.

I think they suggest that 80-90% of the meaning in a face to face human verbal communication comes from non verbal communications like expression and body language.

 

[Fruitloop]

Absolutely. I do however think that there's something unique that happens with the devlopment of genuinely discursive communication like abstract speech or writing - suddenly there are a vast range of much more abstract topics that you can converse about.

 

[fudgefactorfive]

Another point is that language inside the brain is necessarily different to language as expressed in sound or writing. Spoken language has to be linear, there has to be an order of words (interestingly, sign language doesn't have to be linear in this way), mental language isn't constrained in this way.

British Sign Language does have a partially linear (ie. time-ordered) structure (Subject Object Verb as opposed to English's Subject Verb Object) - but because sign language is not one-dimensional (ie. it takes place in 4D spacetime, not in a 1D stream of wah-wah-wahs), it makes very interesting use of space to convey meaning - and uses most of spoken language's features (such as encapsulation) to do it. It doesn't appear to be linear at first glance but in fact, there is a very rigid grammar and time-linear structure there (or it wouldn't be sign language, it would be mime).

Also, look at Latin, a formerly-spoken language which has massively flexible word order, where all the meaning usually attributed to word order resides in the case suffixes. Word order in Latin is little more than convention IIRC.

There has been some evidence (from studying deaf sign users who have suffered brain damage or strokes) that the origin of the brain signals controlling the arms and hands for peple whose first language is sign does not come from the motor function areas but the so-called "language centres", Broca's and Wernicke's areas IIRC (ie. people who have suffered loss of motor control can still sign, although they have trouble picking up objects). This is why someone who learns sign language as a child has such a head start compared to, say, me, who is fumbling along using his motor functions.

People's own first language tends to shape their assumptions about all language, imo.

 

[fudgefactorfive]

I think we think in symbols - whether these constitute "mentalese" is an interesting question, but a very complicated one - it seems to me that the "symbols" are logically at least as remote from "language" as the offside rule is from the price ITV paid for Shiteface Premier League broadcasting rights, or as remote as the chemistry of unsaturated fats is from the menu descriptions of tempura.

But introspection shows that something is going on pre-language. For example, those of you who've done crosswords will recall the experience of the right answer coming to you "in an instant", followed by a much slower process of expressing to yourself why it is the right answer, in language.

I've just listened to Stephen Fry's Radio4 programme "The Joy Of Gibberish" which I thought was very pertinent to this thread, as it asks directly relevant questions such as "who put the baa in the baa-ba-baa-ba-baa, who put the ram in the ram-a-lam-a-ding-dong".

my belief is that it's actually impossible to separate body and brain on this one. we call ourselves self-aware, and all that means is that we "post-process" the sum of all our awareness in the moment after. i can't remember who i read who said that when men ruefully declare that their penis has a mind of its own, they are stating part of a larger truth.

the case of deaf people who've had a stroke that caused pinpoint damage to their motor functions, but can still move their limbs to sign, demonstrates that language is a faculty, that the movement of every muscle in our bodies is not dependent in some linear way on specific bits of brain machinery, but is a result of the state of the entire network. the opposite case is brain damage like aphasia where language construction is totally distorted and impaired - some feedback loop broken somewhere, perhaps - and the sufferer experiences enormous frustration at not being able to put their thoughts into words. yet they are still clearly having thought - they still have internal concepts, emotions.

our entire bodies with their array of glands and other signalling equipment plays a part in every bit of self-reflection we do. yet the ability to process language is just a tiny cog enmeshed in a vast phase space of possible clocks.

language is like the surface of a pond - literally, in that our minds use language to broadcast symbols out of our bodies to other minds capable to processing them - but it does not constitute the eddies and the flow in the higher-dimensional waters beneath the surface.

some people (see post #77 on this thread) claim that language and self-awareness are the same thing. i don't think so. i think that you can't have language without self-awareness, but you can easily have self-awareness without language. as animals clearly demonstrate imo.

"To do is to be" - Rousseau
"To be is to do" - Sartre
"Doo-bee-doo-bee-doo" - Sinatra

 

[weltweit]

Yes it is possible to argue that language is seperate from self awareness and that we only use it for communication and not for thinking in.

A typical communication process goes like this :

1) meaning of the message to be sent

2) encoding (into language)

3) transmission

4) noise

5) reception

6) decoding (language)

7) received meaning of the communication

 

[maomao]

Also, look at Latin, a formerly-spoken language which has massively flexible word order, where all the meaning usually attributed to word order resides in the case suffixes. Word order in Latin is little more than convention IIRC.

Yes, but there has to be a word order, even if the word order can be changed
because you can't say all the words in a sentence at the same time. Whereas in sign language a particle that would have to be a prefix/suffix/infix can be expressed at the same time as the root of a word. Obviously it still has to be partly linear as you can't express all the words in a sentence at the same time (unless you're an octopus ).

 

[Fruitloop]

Is that not mainly 'cos you have two arms and only one mouth? Many's the time I've regretted that particular limitation.

 

[fudgefactorfive]

Yes, but there has to be a word order, even if the word order can be changed
because you can't say all the words in a sentence at the same time. Whereas in sign language a particle that would have to be a prefix/suffix/infix can be expressed at the same time as the root of a word. Obviously it still has to be partly linear as you can't express all the words in a sentence at the same time (unless you're an octopus ).

ah k, with you

"the reason for time is that it stops everything happening all at once"

e2a, although, still, the order of signs is still important, just as in spoken language - if i'm signing "the man kicks the ball", I still have to sign "MAN (index) BALL KICKS" in that temporal sequence, even though I could sign MAN with one hand and BALL with the other - because the sign position in space conveys no subject-object meaning, that comes from sign order, whereas the position a sign has in the space might mean other things, such as being in the past or the future, inside or outside etc.

although the meaning of the simultaneous "MAN/BALL KICKS" sign would be obvious from context, it's not likely that the ball is kicking the man

 

[Fruitloop]

Theres no at once, it's all relative anyway.

 

[perplexis]

He would have needed to paralyse his Broca's Area as well I think.

Yeah, I don't know the study, but using curare to knock out muscle-movement says nothing about the mental processes driving those movements. Curare blocks acetylcholine receptors at neuro-muscular junctions, which blocks the "go" signal from the neuron (sourced from the brain, and as Fruitloop rightly poitns out, probably originating in Broca's area or a related area of premotot cortex). The actual thinking-in-words-process to motor signal won't be affected.
So this proves nothing other than that the articualtory muscles are not required for thought or internal vocalisation.
If you take out Broca's area then you will certainly find problems with articulation, both at the level of muscle-control and at the "higher" level of internal expression.

<wicked thread by the way weltweit. I'm keen to get stuck in with my cogntiive neuroscience hat on...>

 

[perplexis]

<...>
There has been some evidence (from studying deaf sign users who have suffered brain damage or strokes) that the origin of the brain signals controlling the arms and hands for peple whose first language is sign does not come from the motor function areas but the so-called "language centres", Broca's and Wernicke's areas IIRC (ie. people who have suffered loss of motor control can still sign, although they have trouble picking up objects). This is why someone who learns sign language as a child has such a head start compared to, say, me, who is fumbling along using his motor functions.

It's worth pointing out here that Broca's area is in fact a motor area (well pre-motor), which is responsible (roughly, and not exclusively) for controlling articulatory gestures. In the case of a signing individual the articulatory gestures relate not to the vocalisation apparatus but to the arms/hands+face. A signing person who has suffered a stroke which spares motor areas but selectively damaged Broca's area would certainly suffer from impaired signing.

Broca's area receives a lot of input from Wernicke's area, which itself has been increasingly found to be involved in (again roughly and not exclusively) the mapping of sound onto "meaning" (well, at any rate higher-level linguistic concepts than simply speech sounds).

As a result of the dense interconnections between the two (via one of the biggest fibre-tracts in teh brain the arcuate fasciculus) there has been much focus on teh links between speech perception and speech production.

There is (somewhat outdated, but undergoing modification) theory of speech perception called the Motor Theory. It states that all auditory speech input is decoded to articulatory gestures (i.e. sound goes in and the brain processes it in terms of what mouth, tongue, lip, glottal etc actions it would need to execute in order to reproduce the sound). The theory horribly oversimplifies everything we know about the way the brain processes speech, but it does have some truth in it, and recent research has shown that Broca's area is more involved in the processing of difficult-to-understand (e.g. degraded speech) or difficult-to-produce speech sounds (e.g. foreign phonemes with which someone isn't familiar) than in easy ones.

People's own first language tends to shape their assumptions about all language, imo.

This is a very interesting point, and I think it alone puts paid to any notion that we think in words.
It has been shown that people who grow up truly bilingual have an improved meta-conception of language (i.e. they are less constrained by the grammar of a first language than are monglots, since they have learned from infancy that words and grammar are merely formalities, while the real meaning of language is just tagged by these).
To my mind langauge (by which I mean expressive langauge, which we use to communicate) is not the substrate for thought. Or at least, not for all thought. I certainly find that for some types of problem-solving I will think in words.
Of course we then need to examine what means we have for actually answering this question otehr than our introspection. Sadly, at the moment, very few. The problem with teh introspection method is that even if all thoughts ar ejust incredibly fast language (I dobt this very much) there are still thought processes that we cannot conciously apprehend. So, is that language?
In the same way that a picture can make you feel things non-verbally, surely it's reasonable to expect that internal processes (which, let's remember, are all predicated upon the same neuarl substrates as our perceptions of the outside world) also can function without language.

Or, we coull of course turn around and around a narrow definition of thought, which itself would hinge upon thought being something that can only be expressed linguistically. But then we go into philosophical territory which others are far better qualified to get into than I am.

</cognitive neuroscience mode>

 

[fudgefactorfive]

that is teh interesting, i had kind of forgotten temporarily that vocal chords are muscles

 

[weltweit]

It's worth pointing out here that Broca's area is in fact a motor area (well pre-motor), which is responsible (roughly, and not exclusively) for controlling articulatory gestures. In the case of a signing individual the articulatory gestures relate not to the vocalisation apparatus but to the arms/hands+face. A signing person who has suffered a stroke which spares motor areas but selectively damaged Broca's area would certainly suffer from impaired signing.

So sitting alone in a room with your thoughts is significantly different in terms of the parts of the brain that it uses to entering a room full of people and communicating with them. Are you saying that the same parts of the brain are used for sign language as are used for speech? or are you saying that these use different parts of the brain?

Broca's area receives a lot of input from Wernicke's area, which itself has been increasingly found to be involved in (again roughly and not exclusively) the mapping of sound onto "meaning" (well, at any rate higher-level linguistic concepts than simply speech sounds).

I argued in the OP that we may think in language and at some level I still think we do, there are such things as different level languages in computer speak for example you might write some code in something like a 4th generation language which almost makes sense in english but when you have finished it you might have to encode it into machine code so that a computer could understand it. This is what I think happens with humans in that we often learn something in english because that is how the information is presented to us and we take it into our brains at which point it is encoded into a more base level brain language for storage or use for thinking. At the base level our brains must in some way resemble a computer albeit a parralel rather than sequential one because synapses are just digital switches so at the most basic level we are binary thinkers imho.

There is (somewhat outdated, but undergoing modification) theory of speech perception called the Motor Theory. It states that all auditory speech input is decoded to articulatory gestures (i.e. sound goes in and the brain processes it in terms of what mouth, tongue, lip, glottal etc actions it would need to execute in order to reproduce the sound). The theory horribly oversimplifies everything we know about the way the brain processes speech, but it does have some truth in it, and recent research has shown that Broca's area is more involved in the processing of difficult-to-understand (e.g. degraded speech) or difficult-to-produce speech sounds (e.g. foreign phonemes with which someone isn't familiar) than in easy ones.

So there are language decoding areas in the brain? parts that seem to focus on that aspect? or should I say there are language translating areas in the brain because the brain must communicate with itself in some code or other and language is just a code no?

......... In the same way that a picture can make you feel things non-verbally, surely it's reasonable to expect that internal processes (which, let's remember, are all predicated upon the same neuarl substrates as our perceptions of the outside world) also can function without language.

Oh I do think the brain has a language of its own and I think it is probably binary.

Or, we coull of course turn around and around a narrow definition of thought, which itself would hinge upon thought being something that can only be expressed linguistically. But then we go into philosophical territory which others are far better qualified to get into than I am.
</cognitive neuroscience mode>

Thought being something that can only be expressed linguistically?

I think that is possible because expression is to send a message and however we send a message we always encode it into a form of language (even if that is just a facial expression) that we think our intended recipient will be able to understand.

But we can have thought without expression and it may therefore remain in the brains machine code because we have no need to translate it into anything to express it.

 

[laptop]

there are such things as different level languages in computer speak for example you might write some code in something like a 4th generation language which almost makes sense in english but when you have finished it you might have to encode it into machine code so that a computer could understand it. This is what I think happens with humans in that we often learn something in english because that is how the information is presented to us and we take it into our brains at which point it is encoded into a more base level brain language for storage or use for thinking.

Ermm... maybe, for some very abstract definition of "language".

if you start seriously thinking of the brain as a computer - a sequential computer with an instruction set - you will go seriously awry.

You'd do better with an image (at least) 1,000,000,000 smallish machines - routers, say - each yelling at (at least) 10,000 of its neigbours.

So your "lower-level" representations are patterns in that cacophony. It's very unlikely they'd have anything non-philosophers would recognise as a "grammar", let alone a generative grammar.

 

[Fruitloop]

I basically agree totally with what laptop is saying, although it's worth noting that modern computers as opposed to Turing's definition are more like the proposed neural models anyway - multiple objects interact constantly with multiple other ones, and it's only by reference to higher-order informational structures that you know that the contents of this or that area of memory are bits of instructional code, fragments of images or sound etc etc.

Also, a lot of processing is being moved entirely into different modular components meaning that a central processor is only 'aware' in a monitoring sense of functions whose implementation is actually happening somewhere physically separate.

 

[laptop]

modern computers as opposed to Turing's definition are more like the proposed neural models anyway - multiple objects interact constantly with multiple other ones

Erm, but:

• More is different. And 1,000,000,000 is a lot more than 10 or 100.
• Each of the few is a sequential computer with an instruction set.

 

[fudgefactorfive]

... synapses are just digital switches so at the most basic level we are binary thinkers imho.

I'm totally unqualified to say this but as far as I know that is not an accurate description, it's misleading to compare a synapse to a "switch" that is either on or off.

the synapse is the gap between neurons - it's like a connection which sporadically passes information across in the form of neurotransmitters. it doesn't "stay on" or "stay off" like binary computer memory does. it fires, or doesn't fire. cascades of activity are therefore propelled around the network. it's misleading to describes this activity as "digital". the "state" doesn't "rest" in the nodes of the network. it is a constant process of change, entirely in the connections.

 

[ICB]

Good thread, haven't got anything like the time it deserves to do it justice so here's a bullshit overview of where to get stuck in

Given how comparatively basic and problematic neural networks are for those studying AI I'd say the concerns over the computer analogy for the human brain are pretty well justified.

The relationship between language and thought has been a main preoccupation of Analytic (Anglo-American) philosophy since the "linguistic turn" of the mid 20th C. It's also fair to say that it's an incredibly difficult field to get to grips with. Arguably its origins go back through Descartes to Plato, Aristotle and the Stoics (as always).

Wittgenstein and JL Austin are two obvious big names that got it all kicked off in the 40s, moving on from Russell and Frege. A good semi-technical overview is "Groundings in the Philosophy of Language" by Blackburn, his more academic stuff is in "Spreading the Word" which is one of the more highly regarded works on the subject (I can cope with the former, the latter gets beyond me quite quickly.) For those into the logical aspect see also "Meaning, Reference and Necessity", a collection of essays from 1975.

Big complex interplay with the philosophy of logic, epistemology, etc. My head is starting to hurt already.

Big names that I recall from days long gone when I attempted to think about this stuff properly:

William James
Gottlob Frege
PF Strawson
WVO Quine
Alfred Tarski
Saul Kripke
Hilary Putnam
Donald Davidson
Michael Dummet
Dan Dennett
Gilbert Ryle
John Searle
Roger Penrose

Anyone who can explain the relevance of the King of France, Pegasus and the Morning Star wins a prize.

 

[Fruitloop]

Erm, but:

• More is different. And 1,000,000,000 is a lot more than 10 or 100.
• Each of the few is a sequential computer with an instruction set.

In terms of thing one, well yeah, but I also reckon that sometimes more is just more. Also, 10 or 100 is pretty simple for what goes on now - think of a web app where Apache proxies to Tomcat via AJP, to run a java app written with Struts or whatever, which is sitting inside the JVM which is in turn talking to an OS, to device drivers, with monitoring going on of load, power supply, temperatures etc etc. I mean, the brain is pretty mind-boggling in its complexity but in computer terms we're not talking about Moonlander any more. What is the estimated memory capacity of a brain? - about 4 terabytes if memory serves. Pretty huge but looking less outlandish with each passing year.

In terms of the second thing, again I agree completely, but then is the brain somehow not discrete? If not is the continuous aspect essential, or just a side-effect of implementing stuff on luncheon-meat instead of silicon?

 

[weltweit]

...... What is the estimated memory capacity of a brain? - about 4 terabytes if memory serves. Pretty huge but looking less outlandish with each passing year.

Oh that is interesting I had not realised people had estimated that.

I do firmly believe that we have a limited capacity in many ways.

I believe that there is only so much information that you can cram into a brain before information already in there is sacrificed. The memory as a glass of water which can be full after which it just runs over.

However it is also interesting to me that I have a lot of memories of my childhood which I have not accessed for tens of years but I know are still there and only require the right stimulus to become active and accessible again. Sometimes I have wondered if these memories are true complete and whole but there is no way that I can be sure so I just have to accept that they probably are.

I think it is also true that even the most intelligent or well read people only appear to use a fraction of the whole mass that is the brain which makes me wonder what the rest was intended for.

It does not matter if you believe in Darwinian evolution or God like intelligent design, it just appears that there are large volumes of brain matter that are inactive or redundant.

I also think we are limited by a maximum capacity to do things like thinking or acting, we can do a lot and then we are full our capacity has been reached and we just have to stop at maximum.

Are you full yet? have your brains ever been full? mine have, I think.

 

[weltweit]

I'm totally unqualified to say this but as far as I know that is not an accurate description, it's misleading to compare a synapse to a "switch" that is either on or off.

Well it certainly was a simplification I do admit, my understanding is that the processes of the brain seem to be both chemical and electrical but I am no expert I will admit.

the synapse is the gap between neurons - it's like a connection which sporadically passes information across in the form of neurotransmitters. it doesn't "stay on" or "stay off" like binary computer memory does. it fires, or doesn't fire. cascades of activity are therefore propelled around the network. it's misleading to describes this activity as "digital". the "state" doesn't "rest" in the nodes of the network. it is a constant process of change, entirely in the connections.

Well we are simulating this then with factory automation systems now, there are things called bus networks and multiplexing which is basically being able to send lots of different bits of data down the very same wires encoded in a way that they can be decoded at the other end.

If there is a pressure sensor in one building that is sending a permanent signal to its mini control module that it has pressure on it, its constant message gets translated into a pulse of data that is repeatedly sent down the local area network to the central computer which has the task of deciding what to do.

But I guess we don't know if the brain has storage transmission and processing areas or perhaps we do know, I don't know thats for sure hence this thread.

 

[Fruitloop]

Tornadoes aren't simple on-off switches either, and yet they can be modelled by binary programs. The questions is just how much detail you need in order to get an accurate representation (all of it, for complete accuracy in a nonlinear system) and the cost of the computation.

 

[weltweit]

Good thread, haven't got anything like the time it deserves to do it justice so here's a bullshit overview of where to get stuck in

Given how comparatively basic and problematic neural networks are for those studying AI I'd say the concerns over the computer analogy for the human brain are pretty well justified.

I just find it hard to think of a more suitable analogy.

Does anyone have a more suitable one for me?

In electronics you can I think assume something called a "black box" in which you simply list all outputs for all inputs and thus define the functionality of the black box.

It seems far too limiting and simplistic to try to define a brain that way.

There have been studies of children who were brought up by animals who did not get to learn human language early enough in their lives and when reintroduced to humans, could not then learn language and were severely limited by this. What I am meaning is that the total, limited human, inputs into their brains (brain as a black box) limited the outcomes that were available to them. Hope that makes some sense.

Then of course there are almost unique individuals, child prodigies, like Mozart from whom music simply flowed like water from an early age, how could he have learnt all that there is simply no way. In his case outputs vastly exceeded inputs. But these are so far from normal as to almost be able to be ignored.

The relationship between language and thought has been a main preoccupation of Analytic (Anglo-American) philosophy since the "linguistic turn" of the mid 20th C. It's also fair to say that it's an incredibly difficult field to get to grips with. Arguably its origins go back through Descartes to Plato, Aristotle and the Stoics (as always).

Aha, I have not come across any of that. The only thing I have come across (a little only) is NLP (Neuro Linquistic Programming) which seems to come from the USA and which seems to disaprove of psychiatry and is sold as a carrer and life changing technology. I have not been to any courses though and do not plan to.

NLP suggests iirc that you can better control your emotions if you have more tools in the form of a greater vocabulary of words to control yourself with. Thus if you only have "furious" and "livid" as words to describe yourself when you are agitated you may tend to those levels of emotion when even only slightly agitated, because your internal vocabulary gives you little other option. If however you have a whole range of words for being agitated starting perhaps with at the minor end "irritation" then you can better learn to keep control your emotions. I can't claim to know much about it tbh.

Wittgenstein and JL Austin are two obvious big names that got it all kicked off in the 40s, moving on from Russell and Frege. A good semi-technical overview is "Groundings in the Philosophy of Language" by Blackburn, his more academic stuff is in "Spreading the Word" which is one of the more highly regarded works on the subject (I can cope with the former, the latter gets beyond me quite quickly.) For those into the logical aspect see also "Meaning, Reference and Necessity", a collection of essays from 1975.

Big complex interplay with the philosophy of logic, epistemology, etc. My head is starting to hurt already.

Sorry ICB that your head is hurting but if you come back and read this, can you explain what your own view is on the relevance of language to thinking? It would certainly interest me as it seems you have a lot of reading on this from an interesting angle.

Anyone who can explain the relevance of the King of France, Pegasus and the Morning Star wins a prize.

No prize for me then :-(

:-/

 

[118118]

all useful in philosophy illustrations... i have encountered all three in various different settings. if i'm on the right track i will think about it some more, don't think i've ever done a philosophy iq test before lol.

 

[118118]

fwiw i don't think i'm going to remember what i read recently about frege (or whoever). otherwise it could be anything, couldn't it?

 

[maya]

I just read a book written (with the aide of a ghostwriter) by a highly educated autistic woman, who basically claimed that most autistic people think and experience the world non-verbally, i.e. closer to the way animals perceive the world- and that the most seriously afflicted lack any inner dialogue/grasp of language at all...
Apparently for the most high-functioning of them the language ability can be trained to near normal level... She worked as a scientist, but still had some difficulties in communicating in terms of "complex" language-
Interestingly, her research project was looking into communication between autistic people and animals...

 

[Jonti]

Are you talking about Temple Grandin's Animals in Translation, I wonder?

 

[Aldebaran]

Well... then all the born deaf people can't think. Once again a riddle solved.

salaam.

 

[weltweit]

Well... then all the born deaf people can't think. Once again a riddle solved.

hmm that is a good point. Need to think more about that.

Was talking with a buddy about this recently and they thought that they did sometimes think in language but mostly they said they thought they thought in pictures. I suggested they probably thought in mathematics (in which they are quite talented) and they thought that was also a possibility.