Tagging

To analyze an argumentative text or interaction means to tag them according to three main levels.

    1. Delineating the different sequences that compose the text or the interaction under analysis. Characterizing the type and degree of argumentativity of these sequences.
    2. For each argumentative sequence, determining the different lines of argument and their structures; the argument(s), the conclusion(s); the role of the counter-discourses, that is, the kind of mutual criticism and evaluation implemented by each argumentative line.
    3. Specifying the argument schemes.

The analysis of an argumentative sequence must be based on relatively objective criteria, that is to say defined as best as possible, stable and shareable ones, even if not always decisive. The analysis of an argumentative passage is an argumentative activity, whose claims can be criticized and must be justified.

In formal language, one would have markers, that is to say, univocal and automatically identifiable material elements that would allow us to hold discourses such as,

— Presence of marker(s) S: so, this passage is an argumentative sequence.
— Presence of marker(s) A: so, this segment is an argument.
— Presence of marker(s) C: so, this segment is a conclusion.
— Presence of markers T: the underlying argumentation scheme is of such and such a type.

Natural language arguments do not have such markers. Actual linguistic markers are systematically polysemic and polyfunctional. Their strictly argumentative function must be evaluated according to the context. It is as much the context that designates a marker as argumentative as the marker that designates the context as argumentative.

1. Delineating an argumentative sequence

1.1 Sequencing the language flow

At the most general level, if we postulate that language or speech are by nature argumentative the problem of identifying specific argumentative sequences does not arise. If we postulate that, within the macro language datum (text, interaction), only some sequences are more or less argumentative, these sequences must be cut out from the language flow.
A sequence is a relevant analytical unit. The argumentative passages that are exploited in textbooks as well as in scientific presentations, are the product of this first sequencing operation. Sequencing and subsequencing the language flow is, most of the time, a routine operation. The selection of relevant argumentative passages is nonetheless the first problem the analyst must confront, and theirchoice of giving such and such frontiers to the sequence under analysis should therefore be explicit and justified. The correct implementation of this operation supposes a foray into the broader domains of case and corpora building.

In classroom interactions, for example, the sequence, “problem resolution” is distinct from the sequence, “homework and instructions”. In a meeting, the sequence, “agenda setting” is distinct from the sequence “discussion and decision about agenda item 19”. For a participant, identifying the sequence simply means “knowing what we are presently doing”.

Sequences and subsequences of any kind can be defined externally by their boundaries and internally by their own structure and foreground activity.

— Externally, the boundaries of the sequence are transition points characterized by topic changes, by specific closing and opening formulas, and by a re-design of the interaction format.

— From the internal perspective, the sequence is defined by a type of linguistic activity, by a specific interaction format, and by a semantic-thematic coherence, which globally obeys a completion principle. Exactly what a complete sequence consists in depends on the kind of sequence envisaged; the internal principle of completeness of a problem resolution sequence is not the same as that of an agenda setting sequence.

1.2 Delineating argumentative sequences

Argumentative sequences are isolated according to the same internal and external criteria, argumentativity being their specific difference.
Argumentation can be the defining activity of the sequence. It follows that the sequence, “discussion and decision about agenda item n° 19” should normally be highly argumentative. The external boundaries of such an institutionalized argumentative sequence depends upon the rules of the institution.
Argumentation can be an emerging activity in every kind of sequence; for example, somebody disagrees or makes other suggestions during the presentation of the agenda. The left boundary (opening) of an emerging argumentative sequence is characterized by the concretization of an opposition into an argumentative question. The right boundary (closing) can be of any kind, and is sometimes easier to grasp when it is considered in contrast to the following sequence; for example, the Chair looks at the clock and says “Well, I suggest that you further discuss this very interesting point during the coffee break. Thanks for your participation”.

When dealing with a local issue, an argumentative situation may develop and close on the spot, possibly leaving no trace in the memory of the participants.
When dealing with a pre-existing question, such as a socio-scientific issue, the present discussion is just one episode in the larger development of the question as discussed in various settings and crystallized in a specific script. In this case, the question has a story and the sequence is only an episode, which will not close the file.

2. Re-composing argumentative lines: coalitions; argument(s), conclusion(s); criticism and counter-discourse

The internal structure of the argumentative sequence is characterized by the type and the density of the argumentative operations it articulates. Classical analytical points include the following:

— If the text is a multi-speaker interaction or a polyphonic monologue, the analyst must attribute their own to every participant, that is the positions they hold and the roles they play in the global dispute.
Positions are identified as the segment providing, or pointing to, an answer to the argumentative question. Experience shows that this apparently elementary task may be rather challenging.

— Once the exact content of the oppositions and positions has been determined, one can observe the systems of coalition of the relevant positions, as well as their evolution in the dispute.

— The positions being thus localized, one looks at how the surrounding discourse ties with the proposed conclusion-answer, that is to say one identifies the argument(s).
The Indicators of argumentative function, when available, help to identify the passage as an argument or a conclusion.

— The analysis of critical strategies implemented by the participants bears upon the different modalities of counter-discourse management: Direct repetition of other discourses, or evocations, reformulation, of these discourses, rebuttals of the opponents’ arguments. S. Destruction; Refutation; Objection.

— A most interesting point is the observation of the interactions between the the participants, not only the opponent-proponent interactions, but also their interactions with the third parties, and the public at large.

— Observations about the relation between the arguments developed on the spot by the participants and the general script attached to the question, when available, will be always very instructive.

— Further issues involve the global characterization of the argumentative line developed by key participants, as implementing such-and-such strategy.

3. Argument schemes

The argument scheme can be explicitly formulated in the passage as a general law. In order to identify it, one might look for generic statements, which generally function as good supports for the affirmation of values, principles or laws.
To decide upon the correct argumentation scheme, one must investigate whether there is an acceptable paraphrase relationship between the generic discourse corresponding to the argument scheme, and the actual current argumentative discourse under investigation (in classical terms, between the topos and the enthymeme); for a detailed example of this mapping. The operations needed to determine whether such an argument scheme is reflected in the actual argumentation depend heavily on the schema involved. The same concrete argumentative discourse may be matched by several, non-exclusive schemes.