A (simple) definition for philosophy

I like sushi

Personally speaking I think of 'Philosophy' as essentially meaning "ways of thinking about ..." rather than "love of knowledge," which is too question begging for me.

Tarskian

Personally speaking I think of 'Philosophy' as essentially meaning "ways of thinking about ..." rather than "love of knowledge," which is too question begging for me. — I like sushi

One major problem with any definition for philosophy is the following remark:

https://en.m.wikipedia.org/wiki/Definitions_of_philosophy

Another characterization of philosophy sometimes found in the literature is that, at least in principle, it does not take any facts for granted and allows any presupposition to be questioned, including its own methods.[7][11] This is reflected in the fact that philosophy has no solid foundations to build on since whatever foundations one philosopher accepts may be questioned by another.

There is no definition for philosophy.

There is only a collection of partially failed attempts.

The one closest to my own definition of "statements about other statements" is "Thinking about thinking":

https://en.m.wikipedia.org/wiki/Philosophy

Another definition characterizes philosophy as thinking about thinking to emphasize its self-critical, reflective nature.

Overgaard, Søren; Gilbert, Paul; Burwood, Stephen (2013). "What Is Philosophy?". An Introduction to Metaphilosophy. Cambridge University Press. ISBN 978-0-521-19341-2.

Your own proposal of "ways of thinking" is also very close to "thinking about thinking".

I do not look at the process itself -- a black box really -- but only at its output, i.e. statements.

Furthermore, by cutting out the human element, the definition lends itself more readily to computational philosophy.

While a computer can certainly analyze a large body of text, it cannot meaningfully analyze a large number of people while they are actively thinking.

In order to achieve an objective, shared understanding, it is in my opinion preferable to use a reproducible method of analysis.

It may leave out important subsections of philosophy, such as philosophy of the mind, but in my opinion, the benefit of doing so, is greater than the cost.

Ludwig V

All of this originally came up as a remark that the definition for"philosophy" does not need to be computable any more than the definition for "dog" needs to be. — Tarskian

Your project, as far as I understand it, is to develop a machine that can tell the difference between philosophical sentences/statements and other things. I've pointed out that a definition of philosophy may be required for the project, what you describe does not provide it, but rather depends on it.

You may simply have to assume that the promotor of this project knows what he is doing. I don't know anything about dog behaviour, but I would just assume that the project owner does. Since someone else at the Ministry of Health is also willing to pay for the project owner's mistakes, I would give him the benefit of the doubt. — Tarskian

The person in charge of the project to enable machines to identify behavioural and health problems in dogs may well know what they are doing. But they are not developing a definition of "dog".
I can't tell whether you know what you are doing. But it doesn't look good.

Hence, concerning your question, "How is this the relevant to philosophy in any way?", there is your answer, and it is called, "Computational philosophy". It is actually a gigantic subdiscipline. — Tarskian

Well, of course there's a lot of literature and a lot of enthusiasm. Computing is very fashionable. And delegating philosophy to computers might save a lot of wasted time - or be a lot of wasted time. I agree that it is entirely appropriate that the possible applications of computing should be thoroughly explored. But books that you so kindly list for us don't seem likely to provide a definition of philosophy.

Tarskian

But they are not developing a definition of "dog". — Ludwig V

Actually, they are.

It is not in the form of natural language sentences, but in the form of a numerical data structure.

For example, in the following article, they develop a definition for the human face. It is contained in the configuration file haarcascade_frontalface_default.xml. It allows them to do the folllowing:

https://machinelearningmastery.com/using-haar-cascade-for-object-detection/

# Load the Haar cascade for face detection
face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 
             'haarcascade_frontalface_default.xml')
 
# Read the input image
img = cv2.imread(filename)
 
# Convert the image to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
 
# Perform face detection
faces = face_cascade.detectMultiScale(gray, scaleFactor=1.1, 
      minNeighbors=4, minSize=(20, 20))
 
# Draw rectangles around the detected faces
for (x, y, w, h) in faces:
    cv2.rectangle(img, (x, y), (x+w, y+h), (255, 0, 0), 2)

For all intents and purposes, the configuration data in haarcascade_frontalface_default.xml contains an abstract definition for the human face.

At the basis of the classifier is a very, very poor definition of what a visual object is:

Since the technique developed by Paul Viola and Michael Jones in 2001, Haar features and Haar cascades have revolutionized object detection.

Haar features are extracted from rectangular areas in an image. The feature’s value is based on the pixel intensities. Usually, it is calculated using a sliding window, and the area within the window is partitioned into two or more rectangular areas. Haar feature is the difference in the sum of pixel intensities between these areas.

It is believed that an object’s presence will distort the variation of pixel intensity. By checking the pixel intensity between neighboring rectangular areas, you should be able to notice a difference. Hence it is indicative of the object’s presence.

The key idea behind Haar cascade is that only a small number of pixels among the entire image is related to the object in concern. Therefore, it is essential to discard the irrelevant part of the image as quickly as possible.

For machine learning, you do indeed need some notion of what you are looking for. However, it does not need to be perfect at all. Visual object detection is not based on a particularly correct definition of what a visual object is. It is based on a rule of thumb that apparently works out really well.

The presence of Gödel numbering in a sentence, is a notion that is much more directly related to the definition for philosophy than the haar feature is related to a definition for visual object. Using the notion of haar feature in visual object detection works like a charm already. Perfection is absolutely not needed. In fact, perfection is the enemy of good.

Lionino

before I start criticizing them — Ludwig V

And there is the problem. Some things are not up to be criticised because they are evidently correct. It is only by misinterpreting the text, by assuming that the other person is saying something absurd rather than something obvious, fueled by the desire for polemics, that we then enable criticism in something otherwise uncontroversial. Not everything is polemical. In Italy they would just acquiesce and carry on with the conversation, but hereabouts we have "Well, how do you know that you know that you know?".

Lionino

One example for the computability of the term "dog": — Tarskian

First, I said 'dog' doesn't have to be computable, not that it isn't.

Second, the deep learning used to detect dogs can be used to detect philosophical speech, without your distortion of the word.

In fact, go ahead and feed ChatGPT a bunch of texts, it will tell you which is philosophy and which is not.

Inb4: "But some texts are ambiguous and ChatGPT will not get them right!"

Give the dog detector the picture of a dog-looking but alien creature and it will have the same ambiguous and possibility of error, or give it a bad drawing of a dog, or give it a blurred picture of a dog, etc...

Tarskian

In fact, go ahead and feed ChatGPT a bunch of texts, it will tell you which is philosophy and which is not. — Lionino

If you ask ChatGPT about face detection, it will advise you to try OpenCV.

https://medium.com/@vdkolekar/face-detection-implementation-using-chatgpt-622ab3a61df2
(Face Detection implementation using ChatGPT)

ChatGPT cannot do face detection by itself. It will refer you to a specialized engine for visual object detection.

Same with advice on converting PDF to audio. ChatGPT advises to use a specialized engine for PDF (PyPDF2) and one for text-to-speech synthesis (gTTS):

https://medium.com/@vdkolekar/pdf-to-audio-using-chatgpt-ed6d07b98733

ChatGPT does not seem to be itself a specialized engine for anything, actually.

In order to implement philosophy detection based on the definition that I propose, a specialized engine would be required that translates natural English into first-order arithmetic. The only engine I found that could possibly be shoehorned in that direction, is old unsupported abandonware called SPF from Cornell University:

https://github.com/lil-lab/spf

It's not that you can just feed SPF a sentence in English and that SPF will give you the output in the syntax of first-order arithmetic. Instead, SPF looks rather unusable.

Again, there are no budgets for this kind of things. You can much more easily get money for detecting dogs, rabbits, or loose cattle in the field than for philosophy.

Ludwig V

For example, in the following article, they develop a definition for the human face. It is contained in the configuration file haarcascade_frontalface_default.xml . It allows them to do the folllowing: — Tarskian

Sorry, I put my point badly. The definition is being developed by the people, and applied by the machine. So whatever your philosophical machine can do, it is not defining philosophy, but applying a definition of philosophy to the tedious task of distinguishing philosophy texts from other kinds of text.

Ludwig V

It is only by misinterpreting the text, by assuming that the other person is saying something absurd rather than something obvious, fueled by the desire for polemics, that we then enable criticism in something otherwise uncontroversial. — Lionino

That can happen. So it seems a good idea, before embarking on any criticism, to confirm that one's speculations or assumptions are correct or not. No?

Some things are not up to be criticised because they are evidently correct. ..... In Italy they would just acquiesce and carry on with the conversation, — Lionino

Do you never find that something you thought was evidently correct, isn't?
A wild speculation.
Some UK, or at least English-speaking philosophers, especially analytic ones, express disagreement, not by saying "You are talking rubbish" or even "I disagree with you" or even "Why do you believe that", but by saying "I can't understand why you think that.." It's just a politeness - disagreeing without raising the emotional temperature. Perhaps things are not the same in Italy and when you disagree, you say so more plainly.

In this case, what is self-evident to you is not at all self-evident to me. It is a trivial disagreement, since it depends on which criterion you choose to apply count as English or Italian word. Our disagreement is incidental and I didn't intend to provoke a great discussion about it. But perhaps you'll allow me to state my view and then we can leave it at that - agreeing to disagree.

It is true that sometimes similar words and terms are found in different languages, as with "democracy", "'democrazia", "demokratie", "demokracja", "democracia". If "'democrazia" is not an Italian word and not an English word, which language does it belong to? I think you must mean that it is a Greek word. (You'll let me know if that assumption is wrong, I'm sure).
Well, it is not self-evident to me that a word can only ever belong to the language it originally belonged to. If it is meaningful in more than one language, it belongs to all the languages in which it is meaningful. In fact, so this word was originally Greek, but, in various forms (I'll spare you the argument about that) it is now a word in English, Italian, German, Polish and Spanish amongst other languages.

I think it is true of all languages and it is certainly true of English that many words are imported from other languages. Sometimes they remain identifiable as importations, but in many cases they are absorbed. That boundary is pretty uncertain, but I'm not aware that any significant philosophical issues hang on the distinction, so I'm not inclined to worry about it.

Lionino

If you ask ChatGPT about face detection, it will advise you to try OpenCV. — Tarskian

Ok, and?

Lionino

Perhaps things are not the same in Italy and when you disagree, you say so more plainly. — Ludwig V

I am not Italian, just have been and talk to locals frequently.

Do you never find that something you thought was evidently correct, isn't? — Ludwig V

However "'democrazia' non è una parola italiana" is correct, unless one wants to misinterpret it to mean "la parola 'democrazia' non esiste in italiano", which nobody would ever say because it is evidently incorrect.

think you must mean that it is a Greek word. — Ludwig V

And that is because it is a Greek word. Same with 'philosophy'.

Well, it is not self-evident to me that a word can only ever belong to the language it originally belonged to. — Ludwig V

So I recommend that you either look deeper or reconsider your choice of words. It undoubtably belongs to Greek, even if it is used by other languages, because among other things it only makes sense in Greek. In case of disagreement, visit the nearby (authentic) Greek restaurant and tell the restaurant owner that the word 'democracy' "belongs" to English. Even without the same education, his reply will be better than mine would have been.

I think it is true of all languages and it is certainly true of English that many words are imported from other languages — Ludwig V

As long as you arbitrarily choose what qualifies as "many", that is surely true.

but I'm not aware that any significant philosophical issues hang on the distinction, so I'm not inclined to worry about it. — Ludwig V

Though I disagree, my original remark was not philosophical, it was general. This thread itself is not even philosophical. At most, it is a very low-quality attempt at prescriptive grammar.

Ludwig V

Though I disagree, my original remark was not philosophical, it was general. This thread itself is not even philosophical. At most, it is a very low-quality attempt at prescriptive grammar. — Lionino

Yes, indeed. :smile:

Lionino

Though I disagree — Lionino

To be clear, I don't disagree that he is not aware, as obviously I can't read minds, I disagree with the idea that there aren't philosophical issues implied.

jgill

I actually did not invent the term "foundational crisis of mathematics" by myself — Tarskian

OK. Thanks for the links. It should be emphasized that the crises is in the philosophy of mathematics. Mathematicians by and large ignore the crises (unless they are into fundamentals). :cool:

creativesoul

↪Tarskian

Language doesn't think about itself.

Tarskian

Language doesn't think about itself. — creativesoul

Of course, it doesn't.

First-order logic -- in absence of arithmetic -- cannot express sentences about its own sentences. It requires Godel numbering for that purpose which in turn requires arithmetic.

When Godel numbering is available, a person (or a computer) can produce sentences in first-order logic that express sentences that are about other sentences in first-order logic.

A computer does not think but it can output sentences.

Tarskian

This thread itself is not even philosophical. — Lionino

It is about metaphilosophy and computational philosophy based on the philosophy of mathematics.

For you, however, every thread is ultimately always about criticism on people ... except criticism on yourself.

Lionino

It is about metaphilosophy and computational philosophy based on the philosophy of mathematics. — Tarskian

Changing the definition of common words is not metaphilosophy.

And stop crying.

Lionino

There definitely was a foundational crisis in mathematics in the 19th century, it wasn't in the philosophy of mathematics.

One could say the crisis is still going on, as we don't know whether ZFC is free of contradictions (and perhaps never will).

Tarskian

Changing the definition of common words is not metaphilosophy. — Lionino

There is no definition for the term philosophy. There is merely a collection of partially failed attempts. My own definition may be suitable for computational philosophy.

And stop crying. — Lionino

I am several orders better than you at insulting. I just don't do it. I'd rather explore new ideas instead of seeking conflict.

Lionino

There is no definition for the term philosophy. — Tarskian

We went through this already. There are plenty of definitions for philosophy. Not only are you unable to keep up with what has already discussed, but your replies are half of the time off-topic — as you know they are, because you have no proper reply.

The word philosophy doesn't have to be computable any more than the word 'dog' does. — Lionino

The user replies with:

One example for the computability of the term "dog": — Tarskian

As was explained, the point is that it doesn't have to be computable, not that it is not computable. The user does not understand how to appropriately reply to arguments, so by then we are already off-topic.

He uses computability of 'dog' with deep learning as an example.

Second, the deep learning used to detect dogs can be used to detect philosophical speech, without your distortion of the word. — Lionino

By his own example, his proposal for a strict (and erroneous) definition of philosophy is already obsolete, as machine learning can work with the traditional definition of philosophy, and current AIs will show so.

Then,

If you ask ChatGPT about face detection, it will advise you to try OpenCV. — Tarskian

he replies that ChatGPT cannot do face detection, which is completely unrelated and irrelevant (as we weren't talking about ChatGPT and face recognition, but ChatGPT and philosophical texts) and might as well be wrong for next models.

It is really an embarassing discussion. He doesn't even know what it is that he wants to talk about.

I am several orders better than you at insulting. I just don't do it. I'd rather explore new ideas instead of seeking conflict. — Tarskian

It is not an insult. Everytime you find yourself at a dead-end, you start crying about me poking holes in your nonsensical slop. You just keep crying about it, that is the truth. Speaking of insults, you frequently accuse me of being unemployed with zero evidence, even though by your own admission you are jobless in Southeast Asia (among the cheapest civilised places in the world).

Lionino

and might as well be wrong for next models. — Lionino

Not even wrong for "next models", it is already wrong today.

By ChatGPT 3.5o:

Claude nails the image, correctly pointing out China, which is where the picture is from:

All of these are the free edition, I don't pay for any of them.

If you are going to go off-topic, at least don't be wrong about that too.

Tarskian

He uses computability of 'dog' with deep learning as an example. — Lionino

Yes. It points to the existing practice of using specialized engines for object recognition. Music recognition, for example, is carried out with other specialized engines.

Discrimination and classification software is a collection of specialized engines that sport a scripting interface. Computational philosophy most likely needs its own specialized engine.

A suitable engine for computational philosophy does not seem to exist at this point.

me poking holes in your nonsensical slop — Lionino

Maximum precision is a requirement when writing source code. My own source code runs like a charm.

As Linus Torvalds famously quipped, "Talk is cheap. Show me the code."

I don't know about your source code but I doubt that you actually have anything to show for.

When explaining source code, you'd better leave out technical details as much as possible. Source code looks impenetrable to the non-technical person. Hence, less is more.

What I say now, is the result of years of experience of working in a technological environment. You cannot and should not annoy stakeholders with technical details.

Concerning your approach to life, you cannot successfully write functioning source code or do anything of value actually, merely by criticizing other people. As I have asked you previously already, who exactly would want you on their team?

even though by your own admission you are jobless in Southeast Asia — Lionino

If you are really good at writing source code, sooner or later, you won't have to anymore. Successful careers in technology are actually quite short. Next, you just switch to your hobbies instead.

Seriously, do you know of any other ultra high net-worth individual who is looking for a job? I don't look for jobs. Instead, I buy the company if I am interested in it.

Tarskian

you frequently accuse me of being unemployed with zero evidence — Lionino

Someone who never does anything else besides criticizing others, will inevitably have serious problems hanging on to a job.

It's the same situation as with a woman who sleeps around and keeps racking up a growing body count. You just know that she cannot hang on to a husband or even a boyfriend for particularly long. You just know that she is statistically always single.

I assume that you are statistically always unemployed. If you are not right now, then you will soon be.

AmadeusD

You both sound as if it's time to cool your boots, take a couple of days away from each other and come back without the venom present here...

Lionino

Speaking of insults, you frequently accuse me of being unemployed with zero evidence, — Lionino

Someone who never does anything else besides criticizing others, will inevitably have serious problems hanging on to a job. — Tarskian

Like pottery. I am really talking to a badly programmed NPC.

In the meanwhile, it is funny moderators will leave such a post with clownish vitriol and no substance up but erase my post recommending a clearly insane person to seek medication.

Tarskian

In the meanwhile, it is funny moderators will leave such a post with clownish vitriol and no substance up but erase my post recommending a clearly insane person to seek medication. — Lionino

This post is not about you. It is about Godel numbering being indicative of philosophical textual content.

I haven't been able to find a converter of English to first-order logic, but now I have run into something that may reach halfway: grammar parse trees.

https://github.com/opencog/link-grammar

The CMU Link Grammar natural language parser

The Link Grammar Parser exhibits the linguistic (natural language) structure of English, Thai, Russian, Arabic, Persian and limited subsets of a half-dozen other languages. This structure is a graph of typed links (edges) between the words in a sentence.

There is still some distance to cover between this output and first-order logic/arithmetic.

I'm in the gym now but I'll look into compiling the tool later on. There's an entire page on its underlying theory on Wikipedia:

https://en.m.wikipedia.org/wiki/Link_grammar

jgill

One could say the crisis is still going on, as we don't know whether ZFC is free of contradictions (and perhaps never will). — Lionino

I suppose it is, especially among foundations mathematicians. But I would not say it remains a crisis within the broader scope of the profession. Mostly a curiosity.

Tarskian

I suppose it is, especially among foundations mathematicians. But I would not say it remains a crisis within the broader scope of the profession. Mostly a curiosity. — jgill

The discipline of mathematical logic grew out of the foundational crisis:

https://en.m.wikipedia.org/wiki/Foundations_of_mathematics

This led, near the end of the 19th century, to a series of paradoxical mathematical results that challenged the general confidence in reliability and truth of mathematical results. This has been called the foundational crisis of mathematics.

The resolution of this crisis involved the rise of a new mathematical discipline called mathematical logic that includes set theory, model theory, proof theory, computability and computational complexity theory, and more recently, several parts of computer science.

Not sure if mathematical logic is just a curiosity.

jgill

Not sure if mathematical logic is just a curiosity — Tarskian

To make your point would require some sort of poll of mathematicians asking "Are the Foundations of Mathematics important to you as you pursue your explorations into your specialties?"

I'm betting most of my colleagues would say no. I think you are possibly unaware of the enormous scope of mathematical inquiries these days. Look up college catalogues of math courses and find how many have set theory prominently displayed in more than an introductory course. Here are Harvard's Offerings. M145a and M145b and M385 out of how many courses? Plus, of course some overlaps.

When I was somewhat active over 25 years ago Foundations never came up at the conferences I participated in - international groups. Except a joke or two about the continuum hypothesis.

A (simple) definition for philosophy

Welcome to The Philosophy Forum!

Categories

More Discussions