Are Conditionals Really Inferentially Linked Premises?
Comments: 2 - Date: July 29th, 2007 - Categories: Uncategorized
Dr. Tim van Gelder’s excellent recent article on Rationale™ depicts an argument map (p. 7) that can be used to illustrate the question of whether conditionals (e.g. elements of a tort or crime) should be diagrammed like typical “linked” premises that connect through inference steps along a path of reasoning. http://rtnl.files.wordpress.com/2007/07/rationaleforrationale.pdf.
In the above referenced map, the three elements of negligence are depicted in the same manner as inferentially linked premises (e.g. tort element premises and inferential premises are all adjacent to each other.) But in actuality, there are no inference steps that bind them to each other.
I suggest that a more rigorous, effective, and clearer visual grammar with Rationale™ is to separate inferentially linked premises from the conditionals that support them as illustrated below. The horizontal inference path makes clear the inference steps and the vertically placed conditionals reflect their underlying supporting nature. [The premise "the plaintiff was allowed to cross the road unassisted" was moved from the tort element of "plaintiff suffered harm caused by defendant's breach" to the tort element "defendant breached his duty of care." When using a transitive inference structure to link the inferential premises, the more appropriate connection to this different tort element, becomes, in my opinion readily apparent.]
In litigation, I always explain how one “box” is inferentially linked to its neighbor. Factfinders are typically both skeptical and unclear about the nature of inference steps. With a transitive inference pattern, the linkage is obvious. And by separating supporting conditionals from the premises in the direct line of the inferential path, confusion is minimized. [Conditionals still, of course, have probative relevance since they are part of the inferential network.]

Joseph,
A few comments:
- it seems odd to have the main conclusion repeated word-for-word in the primary argument for it. Was this intended?
- by standard Austhink/Rationale conventions, a Reason (green box) should provide evidence *independently* of any other Reasons which may or may not be present and providing evidence for the same point. According to those conventions, this diagram would be malformed since each of three reasons at the second level works *together with* the other reasons so that collectively the provide good evidence for the contention. So, by those same conventions, the constituent premises should all be part of the same Reason.
I realise, of course, that what you are doing is using the Rationale diagrams but providing your own, rather different conventions or “semantics”.
- You are correct, in your comment on my blog, that in the original diagram, the claim “The plaintiff was allowed to cross the road unassisted” seems like it should really be supporting “The defendant breached his duty of care” rather than where I actually (and quite intentionally at the time) put it, as supporting “The plaintiff suffered harm…”. At the time I was attempting to engineer a simple example of an inference objection, and was thinking only about providing an argument for “The plaintiff suffered harm…” which might be subject to an inference objection. When you put the example I came up with in the context of the larger argument, it turns out to be ill-chosen. I’ll have to come up with another, better one.
Hi Tim,
I posted my reply at http://inferencepath.edublogs.org/2007/08/03/designing-an-argument-map-visual-language-for-litigation/ since it has some argument maps.
Joseph
Leave a comment