Politics and Pragmatism in Scientific Ontology Construction

30 Oct 2021 02:15 - 17 Jun 2023 08:29
Open in Logseq
    • In 2011 Carl Hewitt organized a workshop on "Inconsistency Robustness"; I took this as an excuse to indulge my penchant for half-baked philosophical flaming and wrote a paper based on my dissatisfactions with the state of knowledge representation in science. The paper and my slides are available.
    • It was probably my most sustained effort to be a kind of Bruno Latour-ish constructionist and do battle with the naive realism that scientific computation usually assumes.
    • Of course I never bothered to publish this in any place where it would find an audience, and it seems to have been cited exactly once, in a Polish paper:
    • Ten years later and things haven't advanced at all as far as I can tell. I'm still building tooling for class-based ontologies because that is what people understand and know how to work with. The semantic web remains a distant vision; scientific knowledge remains as un-integrated as ever.
    • My own pragmatism has led me to stop pretending to be a philosopher and focus on something I'm actually somewhat qualified to do, that is, building software and bringing what creativity I can muster to the task of bringing people and and the digital realm into a closer working relationship.
    • One sign of hope is that the Roam-led burst of writing tools is starting to merge with semantic-representation. Codex OS is a very interesting project in this space.
    • Pragmatism appears in the title, but am I talking about actual pragmatism or the philosophy? What did I say about it:
      • Philosophical pragmatism was introduced by William James as a corrective to a debate of his own day between materialism and idealism; one that roughly parallels the science wars of today. The essence of pragmatism is to cease looking for the essences of things, but rather to judge them by their consequences: how they are used and how they affect other things. As James put it, “It is astonishing how many philosophical disputes collapse into insignificance the moment that you subject them to this simple test of tracing a concrete consequence.” He further writes:
      • Pragmatism represents a perfectly familiar attitude in philosophy, the empiricist attitude, but. . . in a more radical and in a less objectionable form than it has ever yet assumed. A pragmatist . . . turns away from abstraction and insufficiency. . . .from fixed principles, closed systems, and pretended absolutes and origins. He turns toward concreteness and adequacy, and towards power. . . It means the open air and the possibilities of nature, as against dogma, artificiality, and the pretense of finality in truth.
      • . . . pragmatism [is] a mediator and reconciler. . . that ‘unstiffens’ our theories. . . . pragmatism is willing to take anything, to follow either logic or the senses and to count the humblest and most personal experiences.
        • – William James, [32]
      • Richard Rorty extended pragmatism into the late 20th century and emphasized its antifoundational character [58], and Paul Feyerabend stretched it even further (perhaps too far) with his notion of “epistemological anarchism” [25], an attitude that all forms of representation and inquiry should be allowable. It is just this tendency of the radical pragmatists to take things too far that inspired the realist reaction.
    • And here's where I advanced my own version:
      • 4.5 Summary: What is Representational Pragmatism?

      • Representational pragmatism is an effort to get beyond both the conceptualist and realist factions by positing a different philosophy of what representation is and how it is used. It is not (yet) a concrete technical architecture, although there are many technical systems that in various ways are suggestive (see the next section). Rather than imposing an ontology or even a framework of representation, it takes a computational “anything goes” approach. It encourages lightweight standards, and tools that make it easy to integrate divergent representations.
      • Representational pragmatism is based on what people actually do with representation, in the everyday world and in the real practices of science. It encompasses actual representations found in lab notebooks, articles, whiteboards and everyday discourse as well as the formal structures of ontologists. It is to be found in the processes of formalization rather than in the final formalisms themselves. The role of inconsistency and conflict in this process should be made a strength rather than a weakness. These factors are an inescapable fact of actual representation use, so computational tools should be designed to serve those actual uses, rather than attempt to dictate particular forms of thought. An overly rigid framework for ontology is not robust to the actual structures of knowledge that people need to work with. The next section explores some possible strategies for achieving robustness and flexibility.
    • Goes on to discuss some promising ideas: prototypes, argument and discourse representation, wikis and socially-constructed knowledge bases.