4 ms·

The term "modeling" probably means different things in different research domains. I would expect on the one hand an effort to describe the conjectured dynamics

The term "modeling" probably means different things in different research domains. I would expect on the one hand an effort to describe the conjectured dynamics with some detail and on the other hand a quantification and validation of the model/hypothesis with empirical (historical) data. Admittedly both are extremely challenging tasks: human society is maybe the most complex system we know and we seem very far from adequate modeling in the above sense. This is not to imply that the phenomena identified and discussed are not relevant factors towards a societal "crisis" but they feel more like a partial collection of hunches and analogies which may or may not be capturing the relevant societal dynamics to a meaningful degree. There are reasons to be concerned though. Underlying the theatricality and subjectivity and collective mind games that drive political and economic behaviors there is the physical reality of the anthropocene and its no longer negligible environmental impacts. Its not unreasonable to conjecture that as ad-hoc and capricious social arrangements hit hard external constraints there will be some unavoidable "adjustments". Whether these will unfold as an out-of-control societal collapse or just another chapter of human history remains to be seen. We should not discount the extreme sophistication and resources available to modern societies. Eventually we may even develop a model of society, though its not clear if it will include the impact of there being a model :-)

photonthug 3y ago
> The term "modeling" probably means different things in different research domains. Came here hoping to discuss this although, like many commenters, I skimmed and can’t be bothered to read tfa. Aren’t we at a point with things like stochastic model checkers and process simulation so that stuff like this can be discussed pretty quantitatively? It feels gross that so much erudite economic / sociological analysis that underpins policy decisions isn't more data driven. And I don’t mean data driven as in referencing and then interpreting history, but more like specifying assumptions about agent kinds and their distributions inside simulated worlds of possible policies. Doing economics / decision theory any other way is starting to seem weird, because yeah historically we chased after relatively tidy and closed-form analysis, but now we can just sample the world and then simulate it.
- hackerlight 3y ago
  That does exist, and decision makers do use it, SDGE models for example. But it's of limited utility. Language, rather than math, is often a superior modelling tool for domains where our understanding is inherently fuzzy due to uncertainty caused by a structural lack of data. Sociology would be an example where math is just the wrong tool for the job. If you tried to do that you'd probably be verging on crackpottery like those econophysics people.
  - photonthug 3y ago
    > Sociology would be an example where math is just the wrong tool for the job. really? Stronger statement than just saying the tools aren’t there yet. I also don’t really get the alternative.. surely the phrase “modeling” has connotations of formality and predictive power for any domain, this is desirable, and a “model” should be more of a useful artifact than other subjective interpretations of history (like say an op-ed). Anyway they do mention complex adaptive systems. But whether that means building a model with it, or just nodding vaguely towards it as generally supporting their thesis, I’m not sure
    - xpe 3y ago
      > Anyway they do mention complex adaptive systems. But whether that means building a model with it, or just nodding vaguely towards it as generally supporting their thesis, I’m not sure Done well, models in the complex adaptive systems world can help understand system dynamics at levels better than usual language. Making those models specific enough to make testable predictions is very important but often lacking.
  - xpe 3y ago
    > Language, rather than math, is often a superior modelling tool for domains where our understanding is inherently fuzzy due to uncertainty caused by a structural lack of data. I'll reinterpret your comment charitably: given the current state of how well mathematics is understood, it may well be more effective to use language to reach people. Once someone "checks out" the effort at persuasion has failed. Our culture's lack of mathematical and logical fluency is a weakness. But not all language follows the patterns of rationality, which are formalized with logic and quantified with mathematics. Both have mechanisms for dealing with uncertainty that are far superior to language. I'm not discounting the power of language as a highly-compressed form of information exchange. But when we want to communicate rationality in a compressed way with language, we very often fool ourselves.

photonthug 3y ago

> The term "modeling" probably means different things in different research domains. Came here hoping to discuss this although, like many commenters, I skimmed and can’t be bothered to read tfa. Aren’t we at a point with things like stochastic model checkers and process simulation so that stuff like this can be discussed pretty quantitatively? It feels gross that so much erudite economic / sociological analysis that underpins policy decisions isn't more data driven. And I don’t mean data driven as in referencing and then interpreting history, but more like specifying assumptions about agent kinds and their distributions inside simulated worlds of possible policies. Doing economics / decision theory any other way is starting to seem weird, because yeah historically we chased after relatively tidy and closed-form analysis, but now we can just sample the world and then simulate it.

hackerlight 3y ago
That does exist, and decision makers do use it, SDGE models for example. But it's of limited utility. Language, rather than math, is often a superior modelling tool for domains where our understanding is inherently fuzzy due to uncertainty caused by a structural lack of data. Sociology would be an example where math is just the wrong tool for the job. If you tried to do that you'd probably be verging on crackpottery like those econophysics people.
- photonthug 3y ago
  > Sociology would be an example where math is just the wrong tool for the job. really? Stronger statement than just saying the tools aren’t there yet. I also don’t really get the alternative.. surely the phrase “modeling” has connotations of formality and predictive power for any domain, this is desirable, and a “model” should be more of a useful artifact than other subjective interpretations of history (like say an op-ed). Anyway they do mention complex adaptive systems. But whether that means building a model with it, or just nodding vaguely towards it as generally supporting their thesis, I’m not sure
  - xpe 3y ago
    > Anyway they do mention complex adaptive systems. But whether that means building a model with it, or just nodding vaguely towards it as generally supporting their thesis, I’m not sure Done well, models in the complex adaptive systems world can help understand system dynamics at levels better than usual language. Making those models specific enough to make testable predictions is very important but often lacking.
- xpe 3y ago
  > Language, rather than math, is often a superior modelling tool for domains where our understanding is inherently fuzzy due to uncertainty caused by a structural lack of data. I'll reinterpret your comment charitably: given the current state of how well mathematics is understood, it may well be more effective to use language to reach people. Once someone "checks out" the effort at persuasion has failed. Our culture's lack of mathematical and logical fluency is a weakness. But not all language follows the patterns of rationality, which are formalized with logic and quantified with mathematics. Both have mechanisms for dealing with uncertainty that are far superior to language. I'm not discounting the power of language as a highly-compressed form of information exchange. But when we want to communicate rationality in a compressed way with language, we very often fool ourselves.