The Gift of Gab: What Machines Don’t Have
With a purpose in mind, I know my argument will carry some bias. This, however is acceptable, as this post is designed to elicit debate. You could even say that I’m ‘trolling’, but isn’t that what this is all about?
Within the course, a few assumptions seem to have been made; namely in the realm of what cognitive science is–or more generally, what science is at all. It seems we have taken this older sense of ’science’ (i.e. ideas from Descartes’ time) and equated it with science itself. It really is no wonder, as we have been calling this ‘Cartesian Science’ science since we first touched it. This, I think, has placed emphasis on ‘truth’ or ‘fact’ when, in fact, this is not the directive of ’science.’ A more apt explanation of ’science’ places emphasis on validity; not truth. However, for convenience sake, we often refer to ’scientific truth’ in common parlance as it is easier or more useful to deal with a clearly defined ‘quantum’ term (as in, it is definite). Because of this, I think we are missing some of the nuance behind the question of whether or not a machine can ‘think’ or, more precisely, possess thought abilities equivalent to humans.
As an example of what can happen when we make assumptions based on non-exclusive initial conditions, I provide the common example of an obviously flawed syllogism:
1.) Some men are Greek.
2.) Some Greeks are women.
3.) Therefore, some men are women.
This reasoning is flawed because it relies on both treating the first and second statement as exclusive; i.e.
if you are a man, you are Greek, and if you are Greek, you are a man. This can be shortened to ‘You are a man iff (if and only if) you are Greek’ and also
‘You are Greek iff you are a woman’
Following this, I believe reason only allows a statement to become exclusive when it is first definite. Science, however, does not firmly found a statement as definite, but as currently valid, or possibly temporally definite (definite not absolutely, but only while falsifiable and unfalsified). Therefore, we can only consider machines to think under a certain specific definition only valid in specific frames of reference. However, the same can be applied to the ‘other minds’ problem which faced even Descartes, which I can see will obviously detract from (or more aptly, destroy) my first argument, however, it is still here to provide fodder for all of you to devour. Because of this, I lay out the following argument as a fallback. (it’s kinda crappy too, I know)
Science, and thus cognitive science, only serves to describe the subject of study (in this case, the mind), and from this description we can produce a model. This model is also purely based on calculable factors (by which I mean things enumerable). However, sitting in class today, something simple and very common struck me–computers can’t do small talk. Machines cannot simply blabber on with someone in the cafeteria over crazy politics, suddenly realizing they have a train to catch in ten minutes. Sure, some programmer can sit and write some code (or some code written by some programmer can write some code) setting some parameters, finding some random (or quasi-random) number, and running some little piece of pretty scripting (or running some piece of script to run some other script) to tell the machine that it needs to remember to get to the train station, but the machine never really forgot about the train (nor did it ever have to ‘remember’ as it was merely a set of bits in ‘memory’ tripped), and the machine was never really sidetracked; it was just running code, approximating the results of thought. No matter what this reflects on humans, it is truly the flaws and irrationalities of the human mind which will separate human and machine.
Cognitive science may be all fine and dandy, and from a pragmatic standpoint, it may work to say that a machine can think, but there is much beyond pragmatism. Analyzing the mind as nothing but a series of inputs and outputs will give you data, and if you try hard enough, you can synthesize some relation in data that might simulate the function of thought, but I would not say that simulating thought is the same as thinking. And while pragmatically, this makes little or no difference, philosophically it can mean the world.
Go ahead, tear it up!
7 Responses to 'The Gift of Gab: What Machines Don’t Have'
Leave a Reply
You must be logged in to post a comment.
Whenever someone says that computers can’t think because it’s impossible for one to be wrong, I get very mixed feelings. Sorry this is finally bubbling out on your post and I hope you don’t take offense or anything (not that you should), but saying that computers can’t be wrong or they can’t have smalltalk and just remember they had a train isn’t really saying anything bad about the computers to me. It’s just saying that our mind is like a computer, only broken and shitty.
Joel Ferm
8 Oct 08 at 4:13 pm
I think it’s interesting that when something goes wrong with our computers (although it’s usually that we did something wrong or don’t completely understand the programming or usage), we tend to attribute human emotions to said computer. “It’s angry,” I’ve said about my computer after I click too many times on something and it has a delayed response or shuts out of programs. “It doesn’t like me,” is also common, or “Technology and I don’t get along.” So when we don’t understand something our computers do, we project our emotions or complexities on the machine rather than learning how to fix the problem in a mechanical way. It isn’t a better way of dealing with things. It isn’t good.
Calli Leventis
8 Oct 08 at 11:28 pm
Also, I’ve noticed a lot of people’s arguments against strong AI are based in what makes humans different from computers. These qualities–smalltalk, emotions, being wrong, making mistakes, etc–are not necessarily better, they’re just different. And I’m not even admitting that computers are incapable of these things (after all, it all comes down to the hypothetical question of “is it possible?”), all I’m saying is that lacking these qualities makes it no less possible for a machine to be an autonomous, intelligent being.
Whether or not our brains are “broken and shitty” is another issue, probably one of self-esteem.
Calli Leventis
8 Oct 08 at 11:31 pm
I think our treatment of computers as another being, saying it’s angry or that you and the machine don’t get along, is either a psychological self-handicapping that gives us an excuse to not have failed ourselves in some respects. It’s our nature to treat something we don’t entirely understand as a sentient being. I was talking with Dan and he mentioned our hyperactive agent detection device; if we can perceive it as an agent, we will.
Either that, or we say a computer is angry as a way of getting out of acting as logical and algorithmic as computers are by nature.
Joel Ferm
8 Oct 08 at 11:52 pm
While reading your post, it hit me…what computers lack (or at least a different way of suggesting the same concept we keep discussing) is opinion. Small talk falls under the category/genre of opinion. A programmer can orient/program responses,but they would be based on the “author’s” or “creator’s” opinions. For categories such as abortion that there is no absolute right or wrong answer, you need to base your decision on opinion. If you were to ask a computer, is abortion wrong? The response (if any) would be per the “author’s” opinion or rationale.
Lihy E.
13 Oct 08 at 4:18 am
“it is truly the flaws and irrationalities of the human mind which will separate human and machine.”
This comment begs questioning. Who is to determine what is a flaw or what is irrational? I believe that we are “programmed” or in non-technical terms “reared” and develop flaws. These are based on our programmers/parents. The saying, “the apple doesn’t fall far from the tree” references this concept that we aren’t that different from our parents. IF my argument makes sense, then you can’t really say that programming a machine versus programming a human are different processes (but rather, one in the same).
Lihy E.
13 Oct 08 at 4:22 am
@ Calli
Sometimes, when I put my computer on standby, I check to see if the computer is “breathing” wherein the light fades in and out.
I think that as humans, we like to do whatever it takes to make things “relational.” Perhaps, this is the best/only way to understand certain intricacies. If we relate everything to human behavior/human beings, it will make sense…because afterall (the claim or assumption is WE KNOW EVERYTHING HUMAN). However, we constantly make these assumptions about the “mechanical mind” and claim that it is insufficient in “human” terms. Such claims include that a machine can’t understand like humans, or that small talk is impossible, machines are flawless…etc. the list goes on. Any claim about the the mechanical realm assumes a claim on the human realm. I find it interesting that, as of yet, the human brain (claimed to be mind by some) is still a mystery. How can anyone claim that a machine is “incapable” when we don’t know the full extent of what WE are capable of?
Lihy E.
13 Oct 08 at 4:28 am