The following is my reply to Dr. Tim van Gelder’s helpful comment (http://inferencepath.edublogs.org/2007/07/29/are-conditionals-really-inferentially-linked-premises/) to the previous post in this blog. Since I have included argument maps in my reply, I could not post my reply in the comments section beneath his post.

Hi Tim,

It is good to hear from you. I am glad to hear that the inference observation of your paper was helpful. The question in your comment points to a larger issue, namely, what visual conventions are most effective for argument mapping.

I think the answer to that question is context dependent. For example, a visual language that may be most effective for developing critical thinking skills or other educational purposes may not be most appropriate for persuading a judge or jury. So, perhaps, the place to begin in designing a visual language is to consider the purpose of the argument mapping.

From my years of experience in designing and presenting argument maps for litigation (and, of course, from your tutoring), I have observed several requirements for a visual language for litigation that may differ from an educational context.

  1. There is no real opportunity to teach the new visual language to the judge or jury at trial or even in a brief. This means that the visual language must be self-explanatory.
  2. The exposure to the map maybe only be for a minute or two. This means it must be so intuitive that it can be clearly understood at a glance by the reader.
  3. Because it is always used in a context of conflict, the visual language must allow no vaguenss or ambiguity that can be exploited by the opposing party.
  4. Since it will typically represent a relatively complex argument, it must require minimal effort to read to avoid resistance or inattention in the judge or juror.
  5. Since it it always used in a persuasion context, its use must not engender a risk of feelings of incompetence or frustration in the reader.
  6. Since its sole purpose is to persuade, all features should be designed for that purpose alone. Any concerns that might cause compromising detours for ease of production and pedagogical purposes should be bypassed on the path to the goal of persuasion of a novice reader.
  7. Since the readers, including the judge, very likely have no training in informal logic, the visual language needs to require no understanding of “linked” premises and other theoretical features. The connections and inference steps must be based on a common sense understanding of reasoning.
  8. Since the judge or jury think in terms of parties, rather than arguments from a pedagogical standpoint, the visual language must easily distinguish the reasoning of each party.

To meet these design requirements, I believe that are several relevant principles.

  1. The visual design should resemble as closely as possible the visual language with which the reader already has familiarity. The judge and jury are most familiar with the visual convention of sentences strung front to back in a horizontal path that form separate units. The closer the argument mapping can emulate those conventions, while maintaining its unique attributes, the better.
  2. From new research in the metaphorical and embodied nature of thought and reasoning (e.g. http://www.edge.org/3rd_culture/lakoff/lakoff_p1.html), the visual language should harness the advantages of an embodied metaphor for an intuitive understanding of inference as the foundation for the visual language.
  3. The visual language should provide the necessary fencing (e.g. scaffolding) to keep the reader on the inference path to avoid any slippage of meaning or the reader getting lost and stopping or straying off the path.
  4. The visual language should use colors that identify each party.

The following map uses a visual language I developed that I have so far found best meets these requirements and principles for litigation mapping. It is based on the embodied metaphor of the reader walking in the familiar pattern of left to right horizontally along an inference path. Through the use of transitive inference (e.g. the predicate of one premise is the subject of the next), the reader does not need to understand the difficult concept of “linking” premises. As you have pointed out, understanding this concept can be very difficult. “Logicians and argumentation theorists have found this notion hard to pin down.” http://rtnl.files.wordpress.com/2007/07/rationaleforrationale.pdf.

But with interlocking premises using the transitivity of predication, even children understand the connection naturally. And a transitive inference approach always depicts any necessary hidden premises so the reader does not have to guess at the connection. There must be no gaps in the chain of reasoning that the reader must supply. The inference path also has a START (the subject of the main conclusion) and a FINISH (the predicate of the main conclusion) which bound the reasoning path at both ends and avoids the confusion of basis nodes (e.g. grounds). The different supporting nature of conditionals (e.g. elements of a tort) are depicted in a distinctive manner from the inferentially linked premises. They are shown metaphorically as supporting structures for the premise stepping stones along the inference path. If a conditional is destroyed, the stepping stone that relies upon its support is also at risk.

The following maps illustrate the vagueness in linkage that can result for a juror when a transitive inference path is not followed. The connections following the standard visual conventions are sometimes not obvious even to the practiced eye.

I think this inherent problem of vagueness in linkages with the standard visual conventions is well illustrated by the inferential error that existed in the tort argument map in the Rationale for Rationale™ paper. Likely a number of professional eyes read the map at Austhink and at the Law, Probability, & Risk Journal. Yet, the error was not observed. This is obviously not a reflection on the knowledge and skill of the readers. Rather, I believe it is a result of the vagueness that is inherent in “linked” premises which don’t follow the transitive inference template. (Following “rabbit/hands” is not a complete solution.) Once I translated the standard conventional visual language into the transitive inference visual language, (http://inferencepath.edublogs.org/2007/07/29/are-conditionals-really-inferentially-linked-premises/) the linkage problem was obvious.

Even when a transitive inference approach is used within the conventional pyramid visual language, a number of problems, as illustrated below, still exist. The pyramid approach, typically with a main premise and co-premise as building blocks, requires the unnecessary duplication of premises increasing the complexity of the map as shown below. And since there is not clear START and FINISH, the reader does not have a prescribed path to climb the “mountain” of premises. The reader is forced to wander up and down and up and down without a set order of inference steps to follow. And the vertical travel is, for Western readers, unfamiliar. All of this increases the cognitive load on the jurors…not a good idea.

The conventional visual language, as shown below, does not create “a mountain out of a mole hill.” But it does create a mountain out of a simple path. It is much easier to guide the reader along the well-marked inference path than helping the reader climb a conventional visual language mountain up and down and up and down with no clear route to follow.

Fortunately, Rationale™ is easily adaptable to different visual conventions. I use it for my drafts since it is so user-friendly. I make many changes in my drafts for litigation, and Rationale™ makes that simple and fast.

I want to be clear that adjusting the wording of the premises to fit within a transitive inference path is not always easy. But it can always be done for a structurally valid argument. And for litigation, I want to do the hard work so the judge and jury don’t have to. Its been my experience that if that demand is placed on them by using an unfamiliar visual language, they won’t make the effort.

Best regards,

Joseph