Quick followup idea: we could also treat decision, dissenting opinion, etc. as relations.

Yep, "case" is probably not a good term to use generally speaking. A "case" may have several "decisions" rendered by various courts during it's life.

OK, but so how should we model this? What are the conceptual objects, and what are the relations between them?

I wouldn't distinguish "dissenting opinion" either since it is normally part of the "decision" and is not a separate document really.

But what If I in fact want to cite a particular dissenting opinion; say this one by Scalia? I'm not a legal professional; just a scholar who sometimes uses legal texts.

A lawyer would cite the Scalia dissent using a jumpcite ie: Lawrence v. Texas, 539 U.S. 558, 582 (2003).

I think a lawyer would cite that as:

Lawrence v. Texas, 539 U.S. 558, 582 (2003) (Scalia, J., dissenting).

It's not clear, but I'm guessing you're talking about developing an ontology for citing legal sources in non-legal research. Is that true?

Lawyers, I think, would generalize sources into four broad categories: cases, statutes, journal articles, secondary reference materials. A case is a case. We cite to decisions in cases. There may be various decisions in the life of a case, and each will have its associated documents, primarily briefs, but we make all references in terms of the decision.

Sometimes for important cases that have many decisions, we'll number them, e.g. Handschu I, Handschu II, and so on. But those are like nicknames we use as shorthand to refer to the decisions, and we use them in citations like:

Handschu v. Special Servs. Div., 349 F. Supp. 766 (S.D.N.Y. 1972) (Handschu I)

Whereas you could probably scrape the dissent info from Lexis or another source, I don't think you could do it for the short name.

It's not clear, but I'm guessing you're talking about developing an ontology for citing legal sources in non-legal research. Is that true?

Yes, but one that can be useful for legal people too, or least extended to do so (RDF is designed for extension). So we want the basics right, and perhaps leave details to others.

I talked to someone I know who used to work for Lexis-Nexis, and his argument was it"s best to just focus on documents. The example I gave him (which he seemed to like) was:


<http://ex.net/1> a bibo:LegalDocument ;
dcterms:title "Brown v. Board of Education" ;
dcterms:issued "1954-05-17" ;
bibo:argued "1952-12-09" ;
bibo:argued "1953-12-08" ;
bibo:court <http://ex.net/courts/uss> ;
bibo:volume "247" ;
bibo:pageStart "483" ;
bibo:reverses <http://ex.net/2> ;
dcterms:isPartOf <http://ex.net/3> .

<http://ex.net/3> a bibo:CourtReporter ;
bibo:shortTitle "U.S." .


This is RDF, so you have to visualize a graph. So imagine:


<http://ex.net/1> --- bibo:reverses ---> <http://ex.net/2>


I'm now thinking it might be better to have a subclass of LegalDocument called LegalDecision (along with LegalBrief; we already have Statute).

Does that sound sensible?

First, at least on Lexis, you could scrape the dissent info, but I don't think that's true for concurrences.

As for showing the relationships, have you ever looked at a Shepard's? It's been a while since I've used the book form, but I wonder if it would be useful to look at. Next time I'm in the law library I'll take a look. If I recall correctly, it has a case citation, and then under it a list of all subsequent citing cases.

(Sorry if this is stuff you already know). Shepard's online shows all prior and subsequent rulings in a case. Then it will also show all cases and journal articles that cite to the case. It specifies if a subsequent case overrules, distinguishes, questions, affirms the holding in the subject case. I'm not sure whether it will show a case that has overturned the subject case even if the subject case is not explicitly cited. Also, sometimes a case is overruled in part. And another problem, often a decision will overrule a whole bunch of cases. Btw, I'd use "overrule", rather than reverse.

I do agree that starting with LegalDocument makes more sense. Maybe Opinion and Dissent should be subclasses of Decision.

Yes, I think this is the way to do it:


LegalDocument
Decision
Opinion
Concurrence
Dissent


The way I would explain it, though I don't have much authority for this is that the opinion is the explanation of the majority's result in the decision. The dissent is addressed to the opinion. Concurrences agree with the judgment, but disagree with the rationale in the opinion.

See this as an example:

http://www.law.cornell.edu/supct/html/04-108.ZS.html

You're right on bdarcus. Thanks for clarifying. That's not what I intended. What I mean is that opinions, concurrences, and dissents are parts of a decision. How would you represent that relationship in RDF?