Erwin Van Lun - Jan 4, 2012:

Good thinking! I would suspect that we do not ‘think’ in words but in concepts, which reveal in our conciousness as ‘words’ and stored in our memory in words as well.

Word are concepts.

Grammar defines the structure of the relationships between those concepts.

QUESTION : why the single quotes around the word think in your post?  (what sort of semantics was meant?)

Victor Shulist - Jan 4, 2012:

Word are concepts.
Grammar defines the structure of the relationships between those concepts.

QUESTION : why the single quotes around the word think in your post?  (what sort of semantics was meant?)

Words are concepts, but concepts are not words by definition.

Grammar defines the structure of the relationships between words, not concepts.

ANSWER: ‘think’ is largely an unconcious process of combining information, actually done one neuro level by making new connections or reactivating connections between neurons. Words are a way to express the result of or the intermediate steps in a thought process.

“Words are concepts, but concepts are not words by definition.”

And yet, “concept” is a word.

Sometimes models are easier to manipulate than the real thing; like a higher-level programming language frequently makes it easier to do something you want, than coding in assembly language or physically moving vacuum tubes as the first programmers did. I think natural language gives you the same kind of advantages over pure non-linguistic thought: you can do more with pure words, because the abstraction away from emotion opens up possibilities you were so sure were impossible. Symbols can be analyzed, edited, fit together in new ways, and computers can keep a lot more of them in memory than our limited brains can…

“What language did God actually speak?” I’m going with the Jains and Ardha Magadhi Prakrit :)

“In a few hundred years from now we’ll look back to written language as a ‘old style technology’ that has been around for a few thousand years.”

I hope those who still like to can use it, like vinyl’s still around. I want to have choices. I’d like to have books and libraries in my holodeck (as I like to gather fern fronds to sleep on when camping, like an Indian :), and keep the option of communicating by pure text alone. (The person I’m talking to could always use an interface that transformed my text into something they preferred…)

Erwin Van Lun - Jan 4, 2012:

Words are concepts, but concepts are not words by definition.

Ok, I accept that.  So you’re saying that words are a subset of concepts.  That is correct.    So if words are a subset of concept (or ONE (of perhaps many) ways to ‘encode’ concepts), and there are other types of concepts that are not words, then words *ARE* one way to think about concepts, and just because there may be other ways, does NOT negate that words can be used to communicate and process concepts.  A simple analogy—it is like saying there are two ways to get from one point to another point—let’s say walking and flying.    By saying “words are concepts but concepts are not words” (which I agree, yes)  is like saying “to walk means to move from one point to another” but “to go from one point to another is not walking” (which is correct, since there are many ways to get from one point to another).  *BUT*, that does *not negate* the fact that walking *IS* one way to get from one point to another (words ARE one way to communicate AND process concepts).

Erwin Van Lun - Jan 4, 2012:

Grammar defines the structure of the relationships between words, not concepts.

if words are a subset of concepts, then this is false. 
words := subset of concepts
then
Grammar defines the structure of the relationships between words, which are a subset, or ONE of possibly many ways to represent concepts.

You admitted above , directly that “words are concepts”, thus
  “Grammar defines the structure of the relationships between words” ==  “Grammar defines the structure of the relationships between concepts”

The heart of the issue here is that - just because a more general statement is true, it does NOT negate a more specific version of that statement.

Example: Water is a liquid.
“a cup can hold liquid”—true.
since a water is a liquid (like “one way to represent a concept, or a subset of concepts is words)
it follows the the following is also true
“a cup can hold water”

Just because the more general statement is true that water is a liquid and “a cup can hold liquid” is true, it does NOT negate that “a cup can hold water” is true.

So I think that yes, ‘concept’ has a more broad coverage than ‘word’ - but it doesn’t matter.

Think of how numbers are coded in binary in a computer.  The number 9 (in decimal) is a series of voltage levels, in four bits being: +5V, 0V, +5V, 0V (=1001 binary, or 00001001 in a byte).

Now one can argue— “a series of voltage levels in a machine can be a number”....  BUT… “a number is not a series of voltage levels in a computer”, which is of course TRUE,  but that does not negate the fact that those voltage levels CAN represent numbers, and also voltage levels can represent desired math operations on those numbers.  The computer can do all math operations a human can do even though it is representing numbers are a series of voltage levels.  It does not matter that the concept of “number” is more than “series of voltage levels”.

Erwin Van Lun - Jan 4, 2012:

ANSWER: ‘think’ is largely an unconcious process of combining information, actually done one neuro level by making new connections or reactivating connections between neurons. Words are a way to express the result of or the intermediate steps in a thought process.

I don’t know about you, but I’m consious of my thought processes!

Robert Mitchell - Jan 4, 2012:

“Words are concepts, but concepts are not words by definition.”

And yet, “concept” is a word.

Actually you should specify the difference—whether you are talking about:

(a) “concept” as the text you are seeing on this website right now, OR
(b) what “concept” REFERS to.

So I think you meant , when you said

“concept” is a word , is

“The series of letters ‘concept’ is a word’—  TRUE
“What ‘concept’ refers to is a word’—FALSE
“concept” is a word —-  ambiguious .. do you mean ‘concept’ the series of letters on a screen or paper, or do you mean what it refers to?
“a concept is a word”—clearly false.

another example, “horse” the word that you are seeing on this screen right now, as opposed to what horse REFERS to.

So,
“concept” is a word , correct, but
*what-is-refered-to-as”  concept, is not a word. 

If you know C, it is like a variable,  say myvar;  say its an integer “int myvar = 5;”
then…

int myvar= 5;
int*  myvar_p = &myvar;

so, a “word” would be like “myvar_p” (a pointer or reference to the ‘real thing’), and the actual thing refered to by the word would be the variable itself, myvar;

Robert Mitchell - Jan 4, 2012:

Sometimes models are easier to manipulate than the real thing; like a higher-level programming language frequently makes it easier to do something you want, than coding in assembly language or physically moving vacuum tubes as the first programmers did. I think natural language gives you the same kind of advantages over pure non-linguistic thought: you can do more with pure words, because the abstraction away from emotion opens up possibilities you were so sure were impossible. Symbols can be analyzed, edited, fit together in new ways, and computers can keep a lot more of them in memory than our limited brains can…

“What language did God actually speak?” I’m going with the Jains and Ardha Magadhi Prakrit

“In a few hundred years from now we’ll look back to written language as a ‘old style technology’ that has been around for a few thousand years.”

Robert Mitchell - Jan 4, 2012:

I hope those who still like to can use it, like vinyl’s still around. I want to have choices. I’d like to have books and libraries in my holodeck (as I like to gather fern fronds to sleep on when camping, like an Indian , and keep the option of communicating by pure text alone. (The person I’m talking to could always use an interface that transformed my text into something they preferred…)

AI and the movie “The Time Machine”—I love the idea of AI, for sure.  But a world where humans become useless “vegetables”.  For example, the idea that we shouldn’t even both teaching children how to do basic math because calculators are dirt cheap and everywhere is appalling to me.  I hope AI makes not only machines smarter, but humans also, not stupider.  I want to go for a drive an READ the signs myself… in written language.. not have the computer take control and read my mind and just drive me there!

Oh… Erwin.. what language did God speak?  Easy—- “C”
btw.. we all seem to be communicating very complex CONCEPTS here….....using WORDS… forgive me… .but it seems to me that words really can represent concepts.

Victor Shulist - Jan 4, 2012:

btw.. we all seem to be communicating very complex CONCEPTS here….....using WORDS… forgive me… .but it seems to me that words really can represent concepts.

I thought the question was not whether or not concepts can be represented as words—clearly they can—but whether or not we process thoughts using language or we manipulate concepts in some other form. Perhaps some amalgam of visual memories, audio, lingual, etc. all comprise a concept. Do we think* about this concept in some visio-spatial or otherwise sensory way? Or abstractly via language? Or something else? If so, what?

*By ‘think’ I mean whatever mental manipulation we do to concepts to arrive at new or at least other concepts.

EDIT: Why is the italics tag ([ em]) causing things to appear underlined lately?

Actually I was never arguing that humans use something other than language to think, in fact, I’d say we probably do use something other than language, but I feel that language is somehow part of it, perhaps a big part of it.  But that is just my gut feel… i will let the professionals tell us that when they figure it out (neuroscientists).  They have made much progress,  but I’m not sure if they will have the answer in my life time.. no discredit to them of course, the unthinkable complexity of the brain is just that daunting. 

However, the point I’m pressing is that I think that language is a credible, plausible way to pursue AI.  In the same way we didn’t have to wait to figure out exactly how the mind understands math operations in order to develop a way to mimic that behavior in digital circuitry. 

There are verbal and non-verbal languages. As humans, we use both of these together as a means of communication. There are another layers of communication beyond language which is based on chemical messages and even at the quantum level,  matter entanglemeant that we are only now beginning to explore and understand. So when it comes to AI we use verbal and now even visual information to empower computers to communicate. In the future, I envision communication with machines to be more on the quantum level.

Off topic - CR try using [ i] and [/ i] instead of the em that it puts on by default for italics

I am very interested in the level of dicussion here.
I love it, but my work re-started in a frenetical rhythm hehehe
So I’ll just ask you some questions for matter for thought:

If you have a human child, but he/she is deaf, dumb, blind and illiterate.
1-He thinks? Is he/she intelligent?
2 - How?

3- If you take out language of the equation, but you have an animal and it is able to use tools to manipulate food and modify its environment. Does it think? Is it intelligent?

I think the problem is:
Our brain is composed of 3 layers: Repitilian Brain, Limbic Brain and Neocortex.
AI researchers are trying to emulate only the neocortex: learning, language, logics, racionalization…
when we talk about feelings and mood for example, we are not exactely talking about neocortex.
Fear, parental care, reproductive interaction ( for our racional minds: love)... these are all based on Repitilian Brain.

I’ll comeback to discuss it better.

Victor: Agreed!

Steve: Thanks!

Fatima: Great questions! Definitely animals show astounding levels of intelligence, even to the point of tool use. I’d say this qualifies as intelligent thought, especially when these intelligence tests show adaptation to new environments. (But I guess one must be always wary of anthropomorphizing, no matter how tempting.)

I’m thinking in particular of work with crows that showed they were able to assemble new tools to reach food in situations they had never encountered. Even if the task required assembling a tool to get a part they needed for another tool! Now birds (including crows) show more signs of language-like behavior than most animals, but I seriously doubt we can call them language users. At any rate I doubt they are using language representations to think about the stages of what they are trying to accomplish to get their treat.

I wonder though if by thinking in language we make our brains more efficient at problem solving. Even if not necessary for intelligence, it may prove to radically increase intelligence. Interesting to ponder. I sure hope Victor is just being pessimistic and we get some answers to these questions in our lifetimes.

Laura Patterson - Jan 4, 2012:

There are verbal and non-verbal languages. As humans, we use both of these together as a means of communication. There are another layers of communication beyond language which is based on chemical messages and even at the quantum level,  matter entanglemeant that we are only now beginning to explore and understand. So when it comes to AI we use verbal and now even visual information to empower computers to communicate. In the future, I envision communication with machines to be more on the quantum level.

I was with you up until ‘quantum’. Blegh. Yes, we can encode information in “chemical messages”—I mean, that’s basically all of biology right there. But quantum entanglement does not provide a mechanism for transferring information (communicating). Let me explain. Try not to think of entangled objects as two separate objects, but rather one object that is perhaps spread out in space. Thus the ‘state’ of one part of the object depends on the ‘state’ of its other parts.

And what are these ‘states’? In the lab, they are often the spins of matter particles, or the polarization of light particles. Maybe its some momentum state. It could even be the particle’s position in space. So if the ‘state’ is something like spin, well heck, we use that to store information in digital devices already! Why not entangle two particles and control the spin state of one particle to send information to the other? (In the example you used, you could imagine the machine toggling one end and the other in your own brain, or whatever equivalent sci-fi scenario.) But there’s a catch.

This absolutely will not work.

Although the spin state of one entangled particle of a pair does determine the state of the other, you can never use this property to send information. Let’s imagine I have two electrons. Electron ‘spin’ comes in two varieties—it’s quantized. Let’s call those two spin states ‘up’ and ‘down’. And let’s say I set up two electrons such that their spin is entangled. If one electron has an up spin, the other must have a down.

So why can’t I just start setting one electron’s spin how I like it and let my compatriot, who is sitting with the other electron in some distant place, read out what spin states I’m setting? Because in quantum mechanics, ‘entanglement’ means that we are no longer describing two electrons, we are describing one object with a spin that is both up and down. That is, I cannot say whether the spin of each electron is up or down. To know, I must measure the electron.

When I make a measurement on anything, what am I really doing? I’m interacting with it. When you read your computer screen, photons stream from the monitor and into your eye. Any way in which we register the existence of another object requires that some force carrying particle (photons, etc.) interact with both the object and our measurement tool. The act of interacting with the electron gives us an answer—the spin will be measured as either up or down—but we cannot control which result we will get. Let’s say we measure ‘up’. We have no way of telling the result to our compatriot. If our compatriot also made a measurement simultaneously, they would see ‘down’, but have no way of telling us and no way of learning anything from us. We didn’t decide for the result to be ‘up’, it just worked out that way.

After you’ve interacted with an entangled particle, the entangled state is destroyed—that’s it. Boom. Gone. Your electron is just another lone particle, making its way in this crazy quantum world.

Wow, didn’t mean for this post to be so long. Time for bed!

Fatima - I would say yes, Helen Keller is the perfect example no one can deny!

I like to consider intelligence as the *potential* for abstract processing.  Potential, meaning it doesn’t matter whether or not the mind is actually doing any intellectual work.  Like electrical potential- the voltage is there in the AC outlets of your house, whether you plug something into it or not to cause current to flow through your device to make it work, doesn’t matter.  Intelligence is like the voltage , thinking is like the current.  Continuing that, I guess knowledge would be like the build up magnetic field caused by the current flow… sorry…  another one of my passions is electronics - as I’m sure many have noticed in a lot of my bot conversation example involving electronics

I sometimes believe that we spend so much time and effort trying to determine what intelligence IS, and not simply what it does.  I believe what will happen is we will eventually have a system that does pass a Turing Test, does visual recognition, does speech recognition, can move the robot body that it is inside around, can plan, etc etc, in other words everything a human can do, and be indistinguishable to a human, and *STILL* not have a firm understanding or total agreement about what intelligence is.  Won’t that be a funny situation.

Two fundamentally different perspectives on AI:

[A] The ‘purest’, one who demands that the system *BE* something, internall (such has have a mind that is perhaps even made of the same biological nuerons as us, or at least work in the same manner).  These are the people that claim that, even if an AI can pass a Turing Test and *everything* a human can do, it is still not intelligent (these are supporters of the Chinese Room argument)

the functionalists, like me.  Don’t care how it works, just that it does - that you define intelligence not by how it works, or what it is, but only by what it does.  That was the original idea of the Turing Test afterall.

I know a few people that think/act like reptiles.
Now I understand why.

CR, without re-quoting your entire post, I will agree with you on several points. however, I was speaking from a theoretical point of view since we do not totally understand the quantum world only what we have been able to observe. So if a quantum state is determined at the time of observation, then it would be safe to theorize that the matter that comprises our universe is in constant flux and therefore only exists by direct observation?

Biological circuits have a defined purpose, however their paths are determined by the interaction of other biological agents. So like the quantum world, they are in a state of flux and are also undetermined. Communication is not pre-determined as well, only when facilitated by language or other means of entanglement where there is an actual link that exists between them. Where I am going with all of this gibberish is explained with the Unified Field Theory. UFT is the fabric that binds the quantum states and facilitates communication between the particles.