How do philosophers understand intelligence (beyond artificial intelligence)?

Question

Do philosophers have working definitions of 'intelligence'? The Stanford Encyclopedia of Philosophy provides a lot of references, but all of them are related to artificial intelligence and other fields like animal studies, not to the human and philosophical spheres.

Answer 1

Broadly speaking, the analysis of intelligence tends not to be conducted by philosophers who are more devoted to writing about mind than intelligence. For instance, none of the Stanford Encyclopedia of Philosophy, the Internet Encyclopedia of Philosophy, and the Encyclopedia of Philosophy have entries for intelligence (besides AI). Unlike a topic such as Kant or Descartes in which you can find an endless stream of books, philosophical analysis of intelligence seems to be more scant. After some digging, here seem to be some works with a philosophical bent by authors with philosophical credentials:

• Concepts of Intelligence by Peter Lanz (rowman.com)
• The Concept of Intelligence by Thomas Hally (GB)
• Metaphors of Mind by Robert Sternberg (GB)

The folks most interested in writing about intelligence seems to be psychologists of whom there are two prominent theories which might be seen as representative of two approaches to characterizing intelligence, one devoted to an approach using rigorous psychometrics and the other more pluralistic and devoted to cataloging domains of human expertise. The former is the Cattell-Horn Theory and the latter is Multiple Intelligence theory. The former is based on the study of the G-factor which might be understood as a variable to characterize general intelligence which has now been broken into a number of subdomains. The latter is the latest incarnation of seeing human general intelligence built up from subdomains. The former is much more scientific in terms of reproducibility and grounding, and the latter is far more popular among educators, and its author, Howard Gardner, concedes that the scientific rigor of his theory is less pronounced.

Philosophically, the tendency is to characterize intelligence as dispositions (SEP) of the mind, and modern approaches to the mind by AI researchers and professional philosophers openly accept the modularity of the mind (SEP). I would say that those who study intelligence tend to be reductive materialists who don't stake out claims about mind-body dualism (SEP) which is perhaps one of the central preoccupations of those involved in philosophy of mind, and rather accept the dual basis of mind and body for the purposes of exploring and measuring intelligence.

As such, I would say that the philosophy of intelligence is in its infancy in that the current research interests of philosophers of mind and scientists such as psychologists aren't strongly aligned in the same way that AGI is often seen as a fringe position in AI and their topics of discussion fail to overlap.

Answer 2

Your are puzzled that intelligence is mostly discussed in connection with artificial intelligence. But this is telling: Before the advent of AI, or at least clockworks, intelligence and consciousness respectively "mind" in general were inextricably intertwined. Intelligence clearly was present when self-awareness was present, and could not be thought without it.

Only advances in mechanization, in particular in "mechanizing" information, made a mental distinction between the two possible. The intricate clockworks developed since the 17th century inspired the idea of "something" (a mechanism, an automaton, a puppet) that had the outward appearance of being conscious, intelligent or generally human, but possessed none of the intrinsic properties. Offray de La Mettries Essay L‘homme machine put forth a mechanistic world view, and mechanical dolls appeared in literature, like Olimpia in E.T.A. Hoffman's Sandmann.

The culmination of this distinction is the "Chinese Room" gedankenexperiment brought forward by John Searle. It imagines a mechanism or, broader, "system" which is able to converse in Chinese (which implies some problem-solving capacity), while it is internally just a vast but essentially trivial lookup mechanism.

This demonstrates that the two concepts — consciousness/mind versus intelligence — are categorically different. Consciousness is an intrinsic property, an internal state, while intelligence could be called a "performative" property: The ability to solve problems. It is hard to detect, let alone quantify consciousness; it is comparatively easy to measure at least one definition of intelligence, namely the ability to solve intelligence tests.

Since the concept of intelligence as distinct from mind or consciousness was only made possible by automation, it is also connected to the advances of A.I., which is the continuation of mechanization by other means. As Ray Kurzweil observed, skills that AIs master are silently dropped from the criteria list of "true" intelligence, and in an instant, displaying a degree of amnesia that borders on bigotry, people claim they had never really been on it. (The comment thread to this answer in chat is a good self-referential example.) Here is a list of capabilities which well into the 20th century would have been considered unmistakable proof or the highest intelligence:

• Being an excellent chess player: "It's just simple rules and brute force; you should try a game that cannot be penetrated that way, like Go!"
• Being an excellent Go player.
• Reading.
• Speaking dozens of languages fluently, and translating between them.
• Driving a car in unknown places.
• Suggesting the way proteins fold.

The first thing that has thrown this strategy of implicit retreat off the rails is the advent of ChatGPT. The reason is probably not so much the strength of its "core engine": Many of the above achievements are as remarkable. The main reason is that ChatGPT can sit down and successfully take the Turing test because it has a great speech interface. You don't need to program it. Instead, it can readily, as-is, participate in a lot of human activities. It can pass exams of all sorts, write love letters and come up with cooking recipes.

It is chilling to realize that much of our own behavior is not "truly" intelligent but instead relies on internalized patterns. Most of the times, average lawyers, programmers, cooks, drivers, parents, lovers etc. essentially engage in applying patterns to standard situations, much like ChatGPT. In general, ChatGPT shows us that much of what we do in life is rule-based; sparks of inspiration are rare and far in between.

I think that we will have to abandon this strategy of implicit retreat and stand our ground: Above a certain complexity even rule and pattern based behavior is "intelligent". But then we'll have to accept that we are creating "truly" intelligent machines.

Answer 3

Intelligence is looked at by the field of Cognitive Science, if you wish to label this - which is a more or less loose interdisciplinary conglomerate of several sciences "plus" philosophy. Check through their sources, maybe you find something to your liking.

As the other answer and yourself found out, it seems like intelligence in itself does not really lend itself well to purely philosophical discourse; at least I do not really remember hearing or reading much, if anything, about it in the classic or even modern philosophers. It almost looks like the previous generations thought of intelligence as just a relatively normal feature of humans. Philosophy does not explain other biological human features either (i.e., our capability to move, or eat, etc.).

AI is a bit more of a philosophically interesting topic - obviously most if not all of what we have as A"I" today is anything but; what we are experiencing today is that the algorithms are ever more able to seem intelligent. One could say that ChatGPT and other tools pass the Turing Test, they can fool many people into believing that their output was created by a human, and thus created by an intelligent actor. As you are familiar with the AI world yourself, you know how far that is from the truth. But this is a new phenomenon for philosophy and makes it interesting to discuss, for example, how to detect that some agent (or seeming agent) does have "real" intelligence, or whether there actually are aspects of intelligence which are absolutely congruent between humans and other systems (not only AI, but also, for example, large masses of people, cities, nations, and so on and forth; "unfortunately" it is again more in the realm of other sciences, not philosophy, to think about this; as an example, see "Life 3.0" by Max Tegmark).

On the other hand, philosophers or scientists subscribing to computationalism arive from the other side of the fence and try to explain ever more aspects of the brain and mind using the same physical concepts we know from Theoretical Computer Science (in general, from computational systems, not necessarily physical computers).

Answer 4

According to Aristotle, every physical entity has two aspects, matter and form, which relate to potency and act. He described the 'passive intellect' as the faculty that recieves the forms, and our 'active intellect' as our faculty that acts on the recieved forms, to make inferences. I suggest this 'active intellect' (in Greek, nous) is the first definition of intelligence. He also described humans as having three natures, vegatative, sensitive, and intellective, with the first two shared by anims, and the first only shared by plants.

Aristotle's picture was very influential, and the Scholastic philosophers developed from it the model of humans having Five Wits, or cognitive faculties. It's worth mentioning I think the parallel in Buddhist thought where all schools hold the idea of 'sense gates' with associated cognitive realms, and then a sixth mental faculty 'ideation', and a seventh 'reactivity'. In Mahayana (eg Zen, Tibetan) thought they add an Eighth Consciousness, 'storehouse consciousness', and I'd make a strong case this is analogous to the Noosphere, or Memesphere: the domain of information with substrate-independence.

If you go to the etymology of our modern English word Intelligence it is from, inter "between" + legere "choose, pick out, read". And from the Latin intelligentia which had come to mean "understanding, knowledge, power of discerning; art, skill, taste". The sense of it as information recieved especially from spies, is from the late 1500s.

The Intelligence Quotient, is the estimate by tests of a persons g factor, a hypothesised variable that summarizes positive correlations among different cognitive tasks, reflecting the fact that an individual's performance on one type of cognitive task tends to be comparable to that person's performance on other kinds of cognitive tasks. It began use in education, to identify expected cognitive development for a given age.

Research on normal cognitive development shows us milestones, like development of Theory of Mind in children as evidenced by their development around age three of the capacity to lie about their intentions. That may be impaired for people on the Autistic Spectrum, while cognitive capacity with mathematics or computer coding are often higher, which is interesting. We can look at Theory of Mind in animals, like squirrels and jays which are know to use deception when stashing acorns knowing they are being watched. The Mirror Test is thought to indicate the capacity of self-recognition, which may be crucial to intelligence capacity (eg in the Strange Loops model), but the failure of pigs to pass it while being among the small group of species shown to sponteneously use tools in the wild, indicates it may have limitations.

Our capacity to 'see in' to the minds of others, is foundational to our capacity to mimic and learn visually, to moral reasoning (see Is the Categorical Imperative Simply Bad Math? :)), and conceptual thought and language (see The Private Language Argument, and According to the major theories of concepts, where do meanings come from?). The enormous usefulness of this mode of cognition, makes us biased towards it, and towards making it dels of the world which over rely on it. See Is the idea of a causal chain physical (or even scientific)?

So, intelligence has come to mean different things in different contexts. Our active discriminating intellect, our capacity to absorb information plus assessment from spies, a cognitive faculty with correlating benefits in different mental tasks, and something relating to how humans develop especially as they are educated and which contrasts with animals. And then of course, computing turns up, and of course trying to develop artificial Intelligences is going to be a new context, pushing how we think about the topic. I strongly recommend this lecture by physicist Richard Feynman, which talks in very clear terms about what computers are capable of: Hardware Software and Heuristics.

There are many ways of refining or defining what intelligence is, but it's key to recognise which we pick will deend on context. You pick the context of philosophy, so let's look at some examples there that bear on the topic.

"I seem, then, in just this little thing to be wiser than this man at any rate, that what I do not know I do not think I know either." -Socrates, which I'd relate to his saying higher intelligence requires enter into inquiry & discourse, not rely on verities

"ipsa scientia potestas est (knowledge itself is power)" -Francis Bacon, relates knowing capacity to scope of action (see his use of 'power' elsewhere)

“Intelligence is a fixed goal with variable means of achieving it.” - William James, in 'The Principles of Psychology' (a paraphrase I think, though it's widely quoted and referenced in discussions of the topic)

"Several instrumental values can be identified which are convergent in the sense that their attainment would increase the chances of the agent's goal being realized for a wide range of final goals and a wide range of situations, implying that these instrumental values are likely to be pursued by a broad spectrum of situated intelligent agents." -Nick Bostrom's 'instrumental convergence thesis' which is core to how he defines Superintelligence in his book of that name, ie the transferable increase in being able to attain goals requiring intellectual work

Answer 5

Let me tackle that issue from another angle, because reading a bit between the lines, I think you're asking about sentience rather than intelligence.

From What is IQ and does it matter? an article on the One Central Health website:

IQ stands for intelligence quotient and, in short, it is a measure of a person’s reasoning ability.

In other words, an IQ test is supposed to gauge how well someone can use information and logic to answer questions or make predictions.

From an article in Philosophy Now magazine 'Information, Knowledge & Intelligence':

"Unlike belief and knowledge, intelligence is not information: it is a process, or an innate capacity to use information in order to respond to ever-changing requirements. It is a capacity to acquire, adapt, modify, extend and use information in order to solve problems. Therefore, intelligence is the ability to cope with unpredictable circumstances." -Alistair MacFarlane

Thus we can reasonably state that intelligence of an "AI" and an intelligence of a living being can be considered very similar, so much so that your question is more or less pointless, going with the general definition as above.

However - it's "just" similar, because the devil, as usual, is in the details:

"The differences, both qualitative and quantitative, between human and machine agency can be summarised in terms of three gaps corresponding to different levels of agency: a skills gap, a knowledge gap, and a personhood gap. We cannot hope to match machines in terms of the range and accuracy of their perceptions, the speed and power of their calculations, or the delicacy and precision of their manipulations.

(...)

Nor can we hope to match machines in handling intractable masses of data, or in applying processing power to complex formal systems such as mathematics. Computers are better at storing and retrieving knowledge, and at manipulating formal, symbol-based systems like mathematics. There will be an ever-increasing knowledge gap between human and machine.

(...)

However, there are immensely complex information-processing systems that have evolved in the human brain that cannot be replicated in any machine by any process of formal design or experiment, certainly not for decades to come, perhaps not for centuries. The complexity of our brains is vast.

(...)

So there will remain a personhood gap between human and machine that will continue to make human levels of intelligence, emotional insight and ability to handle uncertainty, unavailable to machines. Within any currently conceivable future horizon of prediction, human and machine agency will remain complementary. We will have to learn how to live with them, but they cannot replace us."

-also from Philosophy Now article 'Information, Knowledge & Intelligence', by Alistair MacFarlane

So to answer your question: yes, the philosophers do have "a definition" of intelligence. It's nice and simple. However, nice and simple answers are rarely good ones, so them philosophers struggle mightily with a good definition of intelligence. However, emergence of "AI" has given new boost and caused an extra focus on this question all around, so hopefully - as shown by the second linked article - there will be something truly usable, sooner rather than later.

I know, Sir Alistair George James MacFarlane CBE FRS FRSE is not someone who I'd call a philosopher, but then again - aren't all scientist first and foremost philosophers first (even though a lot of them forgot it)?

Answer 6

Most of things have a history, that mean thay have atleast two stances - before and after.

Intelligence is an art to find differences batwing before and after (dialectic διχοτομία), "power of discerning", taste of the time, ability to see temporary and eternity.

AI - artificial intelligence, ofc it is only following part after Intelligence art, product of someone's Intelligence - realised copy, image, likeness, but not an Intelligence.

Intelligence ist existence in pair with logic, but if logic is about analysis and brain, intelligence is belonging more to a psyhe structure, or this structure of distinctions you can name a psyhe - principles of the experience gaining complex. Intelligence is for experience gaining, logic is for analysis, classification and same. Logic is based on knowledges, intelligence is based on unknowing.

The main difference ai and intelligence, that ai can to know that it unknown something. That mean that ai can't know something too, and all it's answers is fake copies of someone's intelligence(very rare) or more often simple garbage "opinions" compilation. Also ai needs men's intelligence(souls) to separating the wheat from the chaff. So it is simple one more external tool for men needs, smart one, no more.

And at last about tendencias that "bakers don't know what is intelligence" there is 2 reasons for this, first one they skip this part, how they get knowledges, because they really don't know this - how they got some fundamental knowledges like 2+2=4, because intelligence is that they don't know, it is something before knowing... and second they try to forget something to reknow, that is why they like to say that they don't know something, it is a trick to cheat their unconscious. Ai can't "forget" something.

Answer 7

It is simple. Intelligence is the ability to solve problems. "Wisdom" is the ability to solve them well, whereas being "perceptive" is the ability to discern a problem into a solution. It is a type of "perceptual intelligence", but should be treated separately (as a type of pre-made intelligence from prior ages).

That's it -- there's no need to compare to animals or computers.