Erwin Van Lun - Dec 29, 2011:

@Victor: Even humans experience love (or any other emotion) in a total different way. But still, millions of songs and poems are written about this vague concept.

In order for you to know if different people experience something differently, you would have to actually *BE THOSE PEOPLE* !  You are only one person experiencing things—how do you actually know for certain if others are expereiencing differently?  perhaps they are not experencing at all…. ..  before you respond, this would need its own topic !!!

Erwin Van Lun - Dec 29, 2011:

I belief that computers will be able to experience love, as soon as they get spiritual sensors that measure energy waves around humans and models will be available for emotions.

spiritual? so to me, “love” belongs under the heading “feeling” or “emotion” , and “spiritual” is more of a religious concept.  By saying “sensors” you make it sound like emotions are like electro-magnetic waves being sent out by a TV or radio station lol….  thats a big assumption… I think it is probably something that just occurs within the nuerons of our brains…. again… that requires its own topic.

Erwin Van Lun - Dec 29, 2011:

In today’s AI research however, language is key. And I belief that’s not were the breakthrough will come from.

Oh we differ there greatly, I highly believe language is key.  However, I do appreciate that you said *believe*, and not just “That *is* were the breakthrough will come from”
To be more precise, I think that “language based thought” is one promising area for AI research.

Fatima Pereira - Dec 29, 2011:

Sometimes I think that maybe true AI is like the Ecological Emergent Properties.

I *don’t* think that sometimes….....

I think that all the time!

Highly agreed.

Victor Shulist - Jan 2, 2012:

In order for you to know if different people experience something differently, you would have to actually *BE THOSE PEOPLE* !  You are only one person experiencing things—how do you actually know for certain if others are experiencing differently?  perhaps they are not experiencing at all…. ..  before you respond, this would need its own topic !!!

But Victor, if you are experiencing something, and another is experiencing nothing (under the same circumstances) then by definition the two of you are experiencing the same thing differently. And while you cannot prove different experiences directly, you can certainly gauge by reactions whether or not they’re enjoying their experiences (under most conditions). For example, if you have two different people sit in a theater and watch an opera performance, where one person likes opera and the other doesn’t, you can certainly pick out which is which. From this example, you can determine that you don’t need to be “in a person’s shoes” to have an idea of what they’re experiencing. You won’t know everything down to the smallest minute detail, but you will know enough to determine whether or not they enjoyed themselves.

Why not just ask them? Instead of trying to guess based on projection of yourself into their position, why not make use of this great invention that we have, language?

I’m reminded of the joke about Mr. Rogers asking a jazz musician if he thought about rainbows and flowers when he was playing, and the musician responded “Actually I’m thinking mostly about my financial situation.”

Erwin Van Lun - Dec 29, 2011:

Based on these basic observations, the very first observations of a newborn baby, we are starting to cluster
-which areas of colors belong to one object?
-which object makes which sounds?

[...]

I believe this is the key to human intelligence—the ability to find patterns and group observations into objects and actions, then group those events into more general conceptual objects. Words allow the human mind to hold and manipulate more information than it might otherwise be able to. That’s why we often think in words, even when we have no intention of expressing or sharing our thoughts.

Or maybe I’m putting the cart before the horse. Do we think by manipulating language representations of reality, and then use the new results to visualize the consequences of our thought processes? (ie, words first, translating into images/sounds/etc after) Or do we think “pictorially” (via our senses) and then convert this output into language, even if just for ourselves? Maybe we aren’t actually aware, consciously, of the visualizations/machinations behind the words that pop into our minds. After all, doesn’t our brain go through the same language-less processing when we dream?

But overall I’m betting on the former, at least in terms of our higher mental facilities. Especially when one is concerned with conceptual objects—the only real tool for handling them is via language manipulation, using the logical rules we have learned (represented as specific grammatical and lingual constructions) to reach new thoughts, new conclusions. How or whether we choose to visualize this new information is a later procedure. But language allows us to compartmentalize nuanced events/categories of events that would be hard to hold in our minds if we constantly had to be aware of all the detailed information that the language representation (word/phrase/etc) encompasses.

I bet emotions are an exception though. Often our thoughts bring on feelings that we cannot readily identify the origin of—we must backpedal through our thought processes to figure out what triggered the response, and then put that reaction into words. And as our ability to codify more nuanced reasoning for a physical emotional response becomes more advanced (thanks to language) we can provide more and more specialized names for emotions—and declare that those that require advanced reasoning to understand must be uniquely human. Whether we are right in this assessment is another matter entirely.

Anyway, this is all very fun to think about. I hope that advances in neuroscience will shed more and more light on these questions.

Dave Morton - Jan 2, 2012:

Victor Shulist - Jan 2, 2012:

In order for you to know if different people experience something differently, you would have to actually *BE THOSE PEOPLE* !  You are only one person experiencing things—how do you actually know for certain if others are experiencing differently?  perhaps they are not experiencing at all…. ..  before you respond, this would need its own topic !!!

But Victor, if you are experiencing something, and another is experiencing nothing (under the same circumstances) then by definition the two of you are experiencing the same thing differently. [...]

Ha, true. I think that Victor meant that when two people (for whatever reason) each say they are experiencing love/hate/anger/guilt/etc. that we don’t necessarily need to assume they experience those emotions differently from each other. Even if they did, it doesn’t change the end result—how they behave. Just the same as when two people identify something as red, we can assume they see red the same. The important thing is that no matter what they see, they always put the label ‘red’ on it.

Except that sometimes people *don’t* see red the same, as in the case of color blindness. And it might be reasonable to suppose that there might be variance in the number of various color “cones” that each of us has in our eyes, making even normal-sighted people see differently.

And—here’s the key—this can have important consequences. It might make one person excellent at matching wallpaper and carpet, and someone else hopeless at picking out shirts and ties. And once the *method* behind identifying a concept changes the *behavior* of the person, then Victor’s argument falls apart. If a program’s method for performing some computation works well in most consequences, but leads to failure in others that another method would have succeed at, then perhaps we should be concerned about which method constitutes a better approach.

Which is why there are practical reasons to be concerned with mimicking human behavior using the same mechanisms—or representations of the same mechanisms—that humans use. Otherwise one might (will) run into unexpected places where human and computer behavior differs, probably to the detriment of the computer.

However, all this being said, the average chatbot programmer has more practical interests than re-creating a human mind, whatever the Turing test might say on the issue. And one can develop a serviceable product without fussing over capturing the full range of human output.

Laura Patterson - Dec 29, 2011:

I wonder if the evolution of AI will follow a similar path of it’s biological counterpart?

I think it will too. I think these are the evolutionary steps:
1. Computers needs to be independent of us. When the power drops off, they first need to find a power plug or another way to charge itself.
2. Self Healing: Just like all world creatures: a computer needs to be able to heal itself. It needs to have a blue print of itself, being able to evaluate its current status with this blue print. When it’s not congruent, it’s needs to repair itself
3. Improve by creative recombination of concepts and others. In nature, species constantly try random variations and combinations by allowing their DNA to change on every duplication and recombine with mates. Without the need of having computers to ‘mate’, they can still learn from each other.

Erwin - those are great functionalities and I think they should be on the ‘to-do list’ of all serious AI projects.

CR - yes, your interpretation was correct. 

Dave + CR : how they behave? YES, correct, this goes along with my “functional” view of AI.  AI is not about what it -IS-  (electricity, silicon & algorithms as opposed to flesh & blood), but about what it *does*.  Have a look at my earlier post regarding the relationship between how math is done between digitial circuits & ‘meatware’ (our biological nuerons).

Dave - true about watching their behavior.  but knowing if they like/dislike something is a quite a bit more ‘clear cut’ than knowing the subtle differences about how they feel love.

Robert - yes, ask them, but you are getting only a “representation” of what they fell, Not the *actual* feelings themselves.  Two people could have identical feelings but perhaps express them in language differently.

Anyway,this concept was not central to what I was trying to get across.

C R Hunt - Jan 2, 2012:

Do we think by manipulating language representations of reality, and then use the new results to visualize the consequences of our thought processes? (ie, words first, translating into images/sounds/etc after) Or do we think “pictorially” (via our senses) and then convert this output into language, even if just for ourselves? Maybe we aren’t actually aware, consciously, of the visualizations/machinations behind the words that pop into our minds. After all, doesn’t our brain go through the same language-less processing when we dream?

But overall I’m betting on the former, at least in terms of our higher mental facilities. Especially when one is concerned with conceptual objects—the only real tool for handling them is via language manipulation, using the logical rules we have learned (represented as specific grammatical and lingual constructions) to reach new thoughts, new conclusions. How or whether we choose to visualize this new information is a later procedure. But language allows us to compartmentalize nuanced events/categories of events that would be hard to hold in our minds if we constantly had to be aware of all the detailed information that the language representation (word/phrase/etc) encompasses.

I’ve read quite a number of books on intelligence with apes, our ancestors. Frans De Waal is quite famous for his scientific research in this area, for decades. Apes are extremely smart, but still don’t use language. Our brain has more or less the very structure as apes, but we have a cortex: a show brain tissue that is being able to analyze.

To remember, to cluster, to handle, to interact and to behave however, language is totally obsolete. It’s convenient for communication, that’s all.

Erwin Van Lun - Jan 2, 2012:

To remember, to cluster, to handle, to interact and to behave however, language is totally obsolete. It’s convenient for communication, that’s all.

Without language, how does culture remember the past, so that we can “stand on the shoulders of giants” to build better clustering methods, handle moving earth with machines able to do the work of thousands of men, interact with Jeopardy hosts to defeat champions, and behave like a driver on the road? Symbolic language is what makes knowledge transfer between generations orders of magnitude easier.

Turing with his test recognized that communication is key to intelligence. The easiest way of finding out if something is intelligent is to ask it: Why did you do that? How did you do that? Etc.

A recent slashdot article mentioned apes communicating through ipads. They seem to enjoy it (maybe I could ask Koko via sign language if she’s really enjoying it?). But apes didn’t create the technology to allow them to communicate at long distances, which they seem to want to do. To create the technology that enables such long-distance communication as we’re enjoying right now, and to create future technologies that will enable us to pursue our happiness in more convenient ways, we need language…

Erwin Van Lun - Jan 2, 2012:

To remember, to cluster, to handle, to interact and to behave however, language is totally obsolete. It’s convenient for communication, that’s all.

Wrong.  I can use computer languages to enter a script into my computer, which explains how something is done, execute it, and the machine does my bidding. 

Information ( ==language) is used in our bodies, genes , DNA, which are coded information of how to replicate.  Signals sent between nuerons via synapses, or between neurons and muscles, etc. 

Language can take a lot of forms, from voltage levels representing binary digits, to DNA for building blocks of life,  or the information carried in chemicals between biological nuerons.

@Robert: LOL. What language did God actually speak?

Robert Mitchell - Jan 2, 2012:

Without language, how does culture remember the past, so that we can “stand on the shoulders of giants” to build better clustering methods

Written language is a simplyfied model of spoken language, which is far more subttle. Written language used to have the advantage above spoken language, because of the replication opportunities. However, time is changing, and soon all our kids will learn through voice. 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.

Erwin Van Lun - Jan 3, 2012:

Written language is a simplyfied model of spoken language, which is far more subttle. Written language used to have the advantage above spoken language, because of the replication opportunities. However, time is changing, and soon all our kids will learn through voice. 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 don’t know about that. Calling people by phone is faster than shooting an email, and still people overwhelmingly prefer the latter. The problem with spoken language is that it has a time component. This means 1) you have to formulate what you want to say on the fly, to the detriment of clarity and accuracy, and 2) you can’t “skim” what someone else is saying, whereas you can skim a text for important info and to see if it’s worth reading in full. I don’t see written communication going anywhere as long as people have something complex and of value to say.

As for the importance of language, you state your opinion very boldly for a question still so open. If language were simply about communicating with others, then why do we talk to ourselves? Why do we think through scenarios and work out problems in our heads with words? There’s very good reason to think that language allows humans to use their brain more efficiently than without—that is, it actually makes us more intelligent. Heck, assigning words and grammar is probably intimately tied to our advanced ability to group objects and recognize patterns, the very traits you posit make us intelligent.

The role of language in intelligence is a question that can only be answered in time by neuroscience. (And perhaps the occasional case study of feral children?)

C R Hunt - Jan 4, 2012:

Calling people by phone is faster than shooting an email, and still people overwhelmingly prefer the latter. The problem with spoken language is that it has a time component. This means 1) you have to formulate what you want to say on the fly, to the detriment of clarity and accuracy, and 2) you can’t “skim” what someone else is saying, whereas you can skim a text for important info and to see if it’s worth reading in full. I don’t see written communication going anywhere as long as people have something complex and of value to say.

Apart from your reasoning, there is another good reason why people prefer email. That because human beings prefer the role of having more knowledge than others. They prefer to give information above asking for information. Knowledge means ‘power’. Not having knowledge puts you in a subordinate position. That’s what people don’t like, mostly unconsiously.

Written text can be used for abstraction, just like mathetical formulas. For most common day-to-day communication, I do believe that in the future between human and AI chatbots voice will do the job. The chatbots will always take the subordinate position, always trying to help, and never give the human the feeling that (s)he is lacking knowledge.

C R Hunt - Jan 4, 2012:

If language were simply about communicating with others, then why do we talk to ourselves? Why do we think through scenarios and work out problems in our heads with words?

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.

C R Hunt - Jan 4, 2012:

The role of language in intelligence is a question that can only be answered in time by neuroscience.

Totally agree!

Erwin Van Lun - Jan 3, 2012:

@Robert: LOL. What language did God actually speak?

C

C R Hunt - Jan 4, 2012:

Erwin Van Lun - Jan 3, 2012:

The problem with spoken language is that it has a time component. This means 1) you have to formulate what you want to say on the fly, to the detriment of clarity and accuracy, and 2) you can’t “skim” what someone else is saying,......

,......

Yes, it is like the difference between listening to music on a cassette tape versus CD ... those annoying days of having to stop, rewind, fast-forward….. CD gave us ‘random access’ - press 3 to go directly to song 3.    Written language has that ‘random access’ power. 

Also, think of driving down the highway, will signs not be there to simply read, or will they all have to talk?  The idea that written language will disappear is ridiculous.

Erwin is going to come back with “the car will read your mind” and just take you where you want to go lol.