Friday, April 5, 2013

Don't Care about Meaning

Here's a short email exchange with Caryn, on the common mix-up of meaning and relevance, kept for myself and posteriority.

Why are meaning and caring synonymous in most peoples mind?

By 'meaning', do you refer to 'meaningful life' (pervasive feeling of satisfaction), or semiotic reference? And by 'caring', do you refer to 'compassion for others', 'involvement' or 'motivational relevance'?
Well, to be precise, they relate to negation of meaning with the negation of caring. I think caring for the purpose of this discussion any of the three will work for the moment at least. With respect to meaning, I'm definitely not referring to meaningful life, more the semiotic sense.

I also suspect that we need to define "most people".

I think that most people have no opinion on semiotics and have never reflected on semantics. In this, they have the same unadultered relationship to 'caring' as all the other mammals, they chew what the universe throws at them.

If by most people, you mean "most people with two to ten semesters of philosophy education", or a subset of the intersection of  the sets of people we both know plus the editor of Scientific American, which happens to be a very similar thing, there are probably three lines of argument that amount to what you initially propose:

1. Dread of giving up realism. The main argument is that dropping the understanding of meaning as reference to external facts would kill realism (here I agree), which would open up a hell of unconstructive skepticism and solipsism and should therefore be avoided on common sense grounds, and nihilism (i.e. not caring) would happen to be the (undesirable) normative angle of that hell. Of course, this is a naturalistic fallacy, but since the dreaders are looking for arguments in support of realism, not against it, they won't check very hard.
This is quite similar to the famously dumb "atheism destroys morality" argument, and could probably be countered in the same way (i.e. by collecting hard data on the nihilistic tendencies of non-realists, while controlling for the annoyance caused to them by stupid realists).

2. The intersubjectivity argument: If meaning is the result of a process of social agreements, which bootstraps further social agreements, then dropping meaning would impair higher level sociality. If caring involves higher levels of sociality, it might go down the drain. The argument implies that caring is the result of high-level social cognition (which is at least a little debatable), and that high-level social cognition either depends on our choice of philosophical premises (which is absurd) or which enforces these premises retroactively (because we obviously perform high-level social cognition, and thus all its prerequisites must be given). 

3. The pragmatist stance: Motivational relevance gives rise to our constructivist exploration of our environment. If I turn off the motivational system, I might stop looking for meaning. If I take an entirely pragmatist or performative stance towards meaning (i.e. meaning is what I make it to be), then meaning would entail motivation. The inverse of the argument would be quite different (but similar on the surface): If I stop attaching meanings, I lose the object of my motivation, and thus motivation itself becomes unsustainable.

Since I have no realist, intersubjectivist or pragmatist at hand to test my suspicions, nor a decent sample of meaning-caring equaters to check whether they fall indeed into the above categories, I must admit that all of the above is pure speculation.


  1. Joscha,

    My name is Oscar. I found your blog by looking up your name on Google after watching an interview on YouTube. I noticed that a lot of people were commenting to Adam for the Interview, saying that he had done an awesome job, etc; but I think that the one who we should address is you. Thank you very much for doing that Interview. I find it very interesting that you have a Philosophy background and that you´re creating AI. We cannot do everything in life; we have to choose. If I could chose to do what you´re doing I would, and in part that is why I´m sending you this message. I know that when someone is working on something and has been working on that something for a while, when someone else who knows nothing about the subject starts talking about it or makes a suggestion then it can be very annoying because the newbie does not understand the whole that the Other understands. It is the case here, but still I´d like to annoy you by asking a very specific question. Of course, it depends on you to answer it or not, if you consider it important or not. The question is: Has AI researchers in general or you have considered Immanuel Kant on the creation of AI? What do I mean? The idea of a-apriori related to Kant, in which you cannot suppress time nor space in your imagination. For instance, if we humans picture a room and then we imagine it with no walls, then there´s still space and we cannot suppress space no matter how hard we try. The same applies to time. We cannot conceive of a moment when there is not time, because immediately we ask ourselves: what about the time when there was no time? I wonder if this is something that you have taken into consideration when working on the subject of AGI. If you find it relevant or not and to what extent. And another thing that you don´t address in the interview but is quite interesting is that you believe that "our civilization is pretty much doomed". Why is this so? I know that the answer, if there is such, will be affected whether you reply through here or on a private message. If you wish to respond through a private message let me know to give you my email. Since there are others reading the blog your answer might be different. But maybe you consider it important enough to share it broadly.

    Thank you.

    1. "Has AI researchers in general or you have considered Immanuel Kant on the creation of AI?"

      Most AI researchers do not read Kant, I suppose. I consider him to be the Nostradamus of Cognitive Science: when you read him you realize with hindsight how much he got right, but it is hard to use him as a blueprint in advance.

      Also, most people are not aware that he is much closer to analytical philosophy of mind and constructivism than to the phenomenological schools.

      "The idea of a-apriori related to Kant, in which you cannot suppress time nor space in your imagination. (...) We cannot conceive of a moment when there is not time, because immediately we ask ourselves: what about the time when there was no time?"

      It is quite likely that our brains have dedicated circuitry that raises our inclination to conceive of space and time in certain ways (allocentric and egocentric world model, 3D object permanence, spatial and temporal interval logic, grouping of events into situations and episodic frames, circadian rythms etc.). A huge part of our processing of space and time certainly depends on the nature of the sensory data (light traveling in 3D space etc.), i.e. 3D space + time is a statistical aspect of our input data.

      Furthermore, some aspects of our conceptualizations of space and time depend on cultural influences (see the work by Lakoff and Nunez on the subject, for instance).

      These "instinctive" grasps on space and time are no big help when it comes to an analytical understanding, though. In fact, "truly" understanding a moment without time is not more difficult than understanding a moment with time–but the latter is what we are used to. Being used to something does not mean that we understand it!

      On the other hand, picturing the universe as a computational process, for instance as a sequence of states (that might be orthogonal to the macroscopic direction of the time arrow and a multitude of spatial dimensions) is not too difficult. It is only hard to relate this understanding with our everyday experience.

      Understanding space and time in the cosmological sense is extremely interesting, but I think that it is of little concern to AGI (except for epistemology). The tricky thing for AGI is to get the basic mechanisms right, i.e. how to build a system that can handle the apparent macroscopic 3D space + macroscopic time.

    2. "Our civilization is pretty much doomed": Here is a link that sums up many of the concerns:

  2. Joscha,
    I´m writing a book and one of the themes has to do with how in México and other countries the corruption in the Software Firms and how they are organized exploits the Programmer and achieves very little in terms of innovation. I know this might be way out of what you´re interested in, but would you be willing to participate by sharing the organizational experience of a Firm in Germany? It is a subject I know nothing about and that could act as a model for other Firms to follow. In particular, for instance, and I´ll be brief, there is company in México that I´ve worked with that has a System that separates the Client as much as possible from the Programmers. So much that they tell the Programmers they are charging the client 15USD per work hour, when in fact they are charging them 50USD per work hour. Examples like that.
    Thank you.

    1. I am not sure if I can be a big help. At each local point it looks as if you are innovating and building new things (or making existing things slightly better). Macroscopically, it seems that there is a huge liquidity surplus at the top of the wealth pyramid that seeks investment opportunities. A portion of this gets fed into the IT startup lottery and creates an evolutionary niche, in which the Mexicans (or the Eastern Europeans here) constitute the lowest level of the food chain.

      The internet does indeed flatten the economy geography of the world somewhat, so in theory, individual programmers in Mexico could start their own company and hit it big, but they will find it harder to acquire external funding if they rely on that. If they do contractual work for first world companies, they will have to compete with all the world, while first world contractors have the advantage of closer cultural, geographical and social proximity to the feeding tubes.

      The German IT startup economy seems to be a miniature model of the US blueprint. Many of the successful companies do not focus on innovation, though, but copy successful American web applications for the local market. These have a proven business model and provide a clear exit strategy (i.e. the original company will have an interest in acquiring them). Having understood this, German IT investors have created incubators that will churn out copy cat companies as quickly as they can. On the other hand, innovation does happen for the sake of diversity, but the teams and budgets tend to be smaller than in the US.

      IMHO, a startup economy needs: a startup culture that promotes the right mindset; a bunch of IT oriented universities that produce and mingle entrepreneurial programmers; a culture of investors for angel/seed and VC stages; a market for exit strategies that offer investors an opportunity to recoup their investments.