Argumentation Schemes

Argumentation Schemes represent stereotypical patterns of reasoning. Of particular interest are those described in Douglas Walton's book: Argumentation Schemes for Presumptive Reasoning. Walton's schemes are modelled in the AIF vocabulary, and descriptions for a selection of them have been published by ReasoningLab.

A wider interpretation of "stereotypical patterns" might also include characteristic sub-graphs of an argument map that could be used to condense or summarize it. We'll set these pattern schemes aside for the moment and deal here with presumptive schemes.

Schemes for presumptive inference

schemes.xhtml is a working document describing the presumptive schemes that we'll use within eleatics. It is marked up with RDFa that will generate the corresponding AIF.

Assigning presumptive schemes

In an AIF argument map, we may label any S-node as an example of a presumptive scheme by also typing it as some subclass of aif:Presumptive_Inference. We may label some, all or none of the inferences in this way.

For the purpose of critical analysis, we may consider any claim in an argument map as the conclusion of some presumptive scheme, thereby bringing into play any critical questions associated with that scheme. This makes particular sense when critiquing the claims in an argument map that are premises only, and therefore not justified as the conclusions of some inference step in the argument map. We'll call such premises evidence.

The evidence in an argument can be characterized by argumentation scheme. For example, one claim might be an Expert Opinion, and another an Established Rule. As the evidential claim is not a conclusion, there is no S-node on which to hang the argumentation scheme type as a label. We can get round this by considering the argumentation scheme URI as an RDF container, and then associating an evidential claim with a scheme by making its I-node a member of the relevant container. The Fortitude South example demonstrates assigning evidence to schemes in this way.

Applying presumptive schemes

There is value to be had in just considering the labelling of an argument map by presumptive scheme. See, for example, the subjective evaluation of one of the hypotheses in the Origins of the SARS-CoV-2 virus example.

For the purposes of critical analysis we need to go further and construct a meta-argument around each instance of a presumptive scheme. The points to consider are:

• The meta-argument is a critique. It might be for the purposes of eliciting information, or it might represent some quality or regulatory barrier that an argument is expected to clear.
• Criticism of an argument implies at least two agents. Eliciting information implies a dialogue. The agents may be human or machine.
• Any critique of an argument must add rather than subtract value. Its benefits must outweigh its costs, and its costs are measured in how difficult and time consuming it is for all concerned to engage in the process.