Here, a structured discussion (or "semi-formal discussion") refers to
a set of sentences where each sentence that is not the beginning of a
discussion must be connected to at least
one other sentence by an inter-sentential relation, typically
an argumentation relation (e.g.
counter-example) or a
generalization relation (e.g.
summary). This can be seen as a particular supporting structure
to use in the following related domains:
Computer-Supported Collaborative Argumentation,
Argumentation-based Design Rationale,
Brandom's model of discursive practice.
A structured discussion is a very lightweight form of knowledge representation since only relations between sentences have to be explicited (the content of the sentences do not have to be represented; indeed, in many situations, representing this content would be very difficult, time consuming and irrelevant). From a knowledge representation viewpoint, the relations should be declared and organised or partially defined in an ontology, but allowing undeclared relations may be necessary for encouraging people to participate to the discussion. If the relations must be chosen from small predefined ontology, the resulting discussion structures may be normalized (and hence easier to exploit) but some may be biased or superficial (because the provided relations are inadequate or not precise enough for some discussions). From a knowledge representation viewpoint, a comprehensive and well structured specialization hierarchy of intersentential relations (click here for an example, not particularly comprehensive) should be provided and each user should be allowed to insert new relations at the right place. However, in order not to scare the users, almost all argumentation systems or design-oriented hypertext systems propose less than a dozen relations (for design purposes, it seems that such a small number is not too limiting).
A structured discussion is intended to collect and make explicit the (couter-)arguments for a proposition (that is, the content of a sentence: idea, design choice, fact, etc.) and the relations between these (couter-)arguments. Therefore, using structured discussion may have several goals: easing information sharing, retrieval and comparison, limiting redundancies and recurring discussions, permitting more precise and rigorous discussions, evaluating the "acceptation" or "popularity" of each proposition by people (click here for one possible algorithm to do that), easing decision-making and collaborative work, etc. The above cited domains where structured discussions could be used as supporting structures give much more details on their benefits, even though structured discussions do not necessarily have to be used in those domains. Structured discussions only "lead to the truth" in the sense that they lead to precise statements that most people agree with or recognise that no further argument can be given because it is then only "a matter of personal preferences" (e.g., valuating freedom over life). In any case, a structured discussions are efficient way sto explore, deepen and end debates.
The following notes are supposed to be read while another window shows examples (this may for example be done by "middle-clicking" on the following hyperlinks instead of "left-clicking" on them). First see the structured discussion on XML for knowledge representation and then the structured discussion on abortion. See also this structured discussion about some ideas of Hans-Hermann Hoppe and this one about John Locke's arguments for the existence of God.
1) the syntax (and from a person viewpoint, the indentation) is used for
distinguishing three important constructs:
the relation from a statement, the relation on a relation (or more exactly,
the relation from a statement connecting two statements), and the conjunctive
set of statements. These constructs are important for representing structured
discussions even though few argumentation-oriented hypertext systems offer them
is one of the exceptions; see also
this analysis of Toulmin's argumentation structures).
2) A structured discussion cannot be the direct result of an informal discussion but may be the result of a semi-automatic re-organization of informal discussions and may then be refined/updated by further semi-formal discussions.
3) Relations such as "specialization" or "corrective_restriction" may seem odd to use between informal statements but they are essential for checking the updates of the argumentation structures and hence guiding or exploiting them; specialization relations are used in several argumentation systems (for example, the (counter-)arguments for a statement are valid for its specializations and the (counter-)arguments of the specializations are (counter-)examples for their generalizations).
4) Each of the statements can be re-used independently in various structures and hence cannot refer to some other statement implicitely (the keyword
this used below is a shortcut that must be automatically
generated by the system when displaying a structure: the actual statements
do not contain such a shortcut).
5) Unlike sentences in informal documents, here the statements do not systematically begin by a capital letter so that these statements can be more re-usable. For example, if parts of these statements are directly re-used to generate English sentences, the problem of converting the initial uppercase into a lowercase is never an issue and never has to be solved automatically.
6) You are invited to complement this discussion example (using the same format) or create other (beginnings of) discussions below. When adding a relation and/or a sentence, it is important for the exploitation of these argumentation structures that you indicate your username at the end of the sentence, as "pm" has done below (thus, you indicate who is the creator of the relation and/or sentence); you are not required to have and use a Wikipedia username (you may simply invent a short pseudoname) but you are encouraged to. If many discussions are created, they will be moved to separate pages and indexed by the category "Structured Discussion". To permit these discussions to be printed correctly, the lines should have at most 85 characters.