Machlup on abstraction

"Every one of us thinks most exclusively in abstractions, only the degree of abstraction varies." -- "A note on models in economics"

Here we are getting to the crux of the issue, and the subject of my forthcoming paper comparing Hayek and Oakshott. The world, as the British idealists recognized, is a world of ideas, and foremost a world of concrete ideas. Hayek and his followers, such as Machlup and Vernon Smith, failing to recognize this, regard thought as primarily abstract. The world itself is a mess of incomprehensible "stuff" that we somehow understand by forming abstractions. But the idea that we can create comprehensible abstractions from a welter of incomprehensible particulars is logical rubbish. And this fundamental philosophical mistake vitiates Hayek's understanding of rationalism, rendering it inferior to Oakeshott's.

Comments

  1. I strongly disagree. It is not that the particulars are incomprehensible. It is that particulars have many aspects. And we understand particulars only to the degree that we can say it has this aspect, and that aspect, etc. Those aspects are abstractions.

    There are concretes, but we cannot study them except through abstraction.

    ReplyDelete
    Replies
    1. How did you come to form these abstractions, marris? When you were at your mothers breast, Did you only know what to do Because you had the abstraction "mammary gland" in mind? (Sorry about the weird caps: I'm on my phone, so I am dictating: Blogger comments or a nightmare from a phone.)

      Delete
    2. Or Consider the many animals that can differentiate between a pile of food with more items in it from one with fewer. Do you think that they understand numbers?

      Delete
    3. Finally, I wish to comment on your introduction of "study them". If you mean "theorize about them," then I agree: we theorize by forming Abstractions.

      Delete
    4. > How did you come to form these abstractions, marris?

      I dunno. I'm not sure that question is answerable for all minds without a lot of philosophical grand-standing. Maybe we will one day be able to answer the related physical question: what happens to the brain when it applies a concept; if there is firing of neurons in a particular structure, how that structure formed; what structures grow in isolated development, but which ones require your parents to show you stuff as a baby (e.g. hold a ball and say "ball").

      > Consider the many animals that can differentiate between a pile of food with more items in it from one with fewer...

      This is probably an argument *against* the primacy of experience. It is true that some animals are shown how to hunt, etc by their parents. But are there ones which, if you separate them from their parents at birth, can still make this distinction? It is plausible that the ability to go for the larger pile is *wired in* by evolution.

      I'm totally willing to admit that the brain may have higher level concepts and lower level sensory data, but I'm not sure we can compare new sensory data to previously *experienced* sensory data without abstraction. I'm not sure about this.

      Finally, I with to comment on "the idea that we can create comprehensible abstractions from a welter of incomprehensible particulars is logical rubbish".

      That's logical rubbish at all! Consider machine learning, in which we start with a fixed set of features (some structures of the brain that are specified biologically) and calibrate the weights of these features with a data set (our interactions with our environment). As the system learns, it becomes a better model for the world.

      Delete
    5. "I'm not sure that question is answerable for all minds without a lot of philosophical grand-standing."

      I'll answer it for you: you form them by abstracting FROM concretes already understood.

      "It is plausible that the ability to go for the larger pile is *wired in* by evolution."

      So? How does this affect my argument at all?

      "lower level sensory data"

      There is no such thing as "sensory data" that is not already an idea.

      "Consider machine learning"

      Which is just a metaphor.

      You have not even tried to touch the logical point here, just thrown in a red herring metaphor. How can you form a sensible abstraction from a bunch of things that individually make no sense to you?

      Delete
    6. > I'll answer it for you: you form them by abstracting FROM concretes already understood.

      Sorry Gene, assertions don't count. I'm disputing your claim that you can understand concretes at all. I'm saying we have sensory data (subjective experience, etc) and we have abstractions.

      > There is no such thing as "sensory data" that is not already an idea.

      Nope. Animals can detect and respond to light vs. dark, more vs. less, etc without forming *ideas* about them. Similarly, people feel emotions, hot vs. cold, etc without necessarily forming ideas about them. How can we understand reflexes without accepting the existence of sensory data?

      > Which is just a metaphor. You have not even tried to touch the logical point here.

      I'm not using it as a metaphor. I'm using it as an analogy. I think that neurons rewire themselves and change their triggering levels based on sensory data in the same way that an ML model changes it's coefficients as more data comes in.

      Delete
  2. This comment has been removed by a blog administrator.

    ReplyDelete

Post a Comment

Popular posts from this blog

Libertarians, My Libertarians!

"Machine Learning"

"Pre-Galilean" Foolishness