• JosephS
    74
    I come from a background of data extraction, manipulation and analysis (Hive, Neo4j, ML, NLP) and have an interest in a practical, tactical implementation of these techniques to the study of ethics as it applies to law.

    I've found references to automated argument extract and the philosophy of action, but have been unable to find anything tangible to evaluate the feasibility or even coherence of my hobbyist effort.

    Where I'm having difficulty is finding resources regarding formalisms on the philosophy of action. I'm looking specifically at a formalism with an application to law (and/or ethics). The categorization of actions into permissible and impermissible with corresponding sub-categories related to the justification in ethics or law is the end goal.

    If anyone here would mind commenting on resources that I might find useful to my goal, I would appreciate it.

    In as much as I have so far been working in a vacuum, if you want to comment or critique on the sanity (or lack thereof) of the effort, please do.

    Thanks.
  • fdrake
    2.5k
    So let's postpone the is ought gap for a second; studying moral sentiment will not tell you what is moral or immoral, it will tell you what people think is moral or immoral. If you want to make data out of it, you should make something like a questionnaire that poses various permissibility/impermissibility questions and elicits demographic data from the recipients.

    If you actually want to study legal systems, you're not going to get much of an improvement from automated approaches over qualitative and historical comparative methodologies. At best you will be summarising trends in legal documents and rulings.

    You've also got a two level problem, where people are nested within legal systems but have individual level variability. That's a hierarchical modelling problem.
  • fdrake
    2.5k
    You'll need to come up with really clever operationalisation and get a lot of data from loads of legal systems and people in them to say anything relevant at all. It's definitely more of a survey design, data collection and model problem than a "throw a neural network at it and it'll be fine" problem.
  • JosephS
    74
    The is/ought is intended as a much longer-term project (steps 5 or 6, if not 105).

    The immediate interest is understanding the analysis of action as it applies to law.

    The way I've expressed it in my naive fashion is:
    X Does A With C - X is a living human being who commits A, an Action, and C is a the set of Conditions under which A is done. If it is a fact that “Professor Plum murdered Miss Scarlet with the wrench in the conservatory”, then X is “Professor Plum”, A is “hit Miss Scarlet with the wrench” and C includes “Miss Scarlet died as the direct result of A by X” and “A occurred in the conservatory”. C could also include related actions indicative of intent or any other conditions relevant to the determination of whether the principal action is permitted/prohibited.

    Rather than continuing to dwell in the confines of my ignorance, I'm looking to understand how researchers in the philosophy of law and action break this topic of analysis down.

    The nice part, I would think, of this analysis is that the corpus of legal case dispositions provides concrete material to play with. I haven't yet found that public dataset but have communicated with individuals working on them.

    Bringing it up to the level of general ethical principles (rather than specific to US legal principles) will only occur after the portion which analyzes actions of those tried and the justifications for legal action, extracted from a corpus of case results.
  • fdrake
    2.5k
    So you're looking for like a regex search procedure of legal documents to get the data out? Statements like that. I don't know how you'd learn the ontology without assuming some of the structure from the regex calls or data selection criteria.
  • JosephS
    74
    Yeah, I'm still playing with what the operational analysis would look like. Depends on the size of the dataset and how it is structured. tf/idf is something I've used in the past, but I suspect that my strategy on ELT vs ETL and the ML algorithms used will have to be torn up and redrawn several times, if my past experience is any guide.

    In the mind of 'fast fail', I'm trying to figure out if I need to be committed for even contemplating this.

    I've never worked in the legal field. My experience is supply chain, finance and cust sat. I'm starting at ground zero, so I'm looking at trying to understand the subject matter of legal actions from a philosophical/analytical base. My hope is that understanding how actions are parsed will enlighten me on how best to appreciate the landscape of legally permitted/prohibited actions.
  • JosephS
    74
    Here's a link to a news item regarding a dataset currently under construction.
    https://www.law.ox.ac.uk/business-law-blog/blog/2018/05/law-and-autonomous-systems-series-paving-way-legal-artificial

    The content (100k cases) and structure it promises would be a boon to my effort. I've looked for similar public datasets but haven't come across any.
  • fdrake
    2.5k
    Interesting.

    Terrifying potential for oversight and fake objectivity.
  • Galuchat
    625
    I'm looking specifically at a formalism with an application to law (and/or ethics). The categorization of actions into permissible and impermissible with corresponding sub-categories related to the justification in ethics or law is the end goal.JosephS

    See Robert Schirokauer Hartman's scientific axiology. Marvin Charles Katz ("prof" on the old forum) was one of his students.

    I think that the quantification of ethics (morality/immorality) is possible, and that its application in the formation of public policy is appropriate.
bold
italic
underline
strike
code
quote
ulist
image
url
mention
reveal
youtube
tweet
Add a Comment

Welcome to The Philosophy Forum!

Get involved in philosophical discussions about knowledge, truth, language, consciousness, science, politics, religion, logic and mathematics, art, history, and lots more. No ads, no clutter, and very little agreement — just fascinating conversations.