On the Norms of Visual Argument: A Case for Normative Non-revisionism

Following Dove (2011: 4) I will take a visual argument to be “any argument in which at least one of the elements is conveyed visually” (cf. Blair 2015: 218). As such, included in the category of visual arguments, as they are discussed in the paper, are arguments of mixed-modality, or multimodal arguments, where at least one of the elements of the argument is conveyed visually.

Here I employ, with a neutral connotation, the language of the “innovation lifecycle”—a sociological model intended to map the adoption of some change (frequently technological) within some group. The model, known as the diffusion process (Bohlen and Beal 1957; Beal et al. 1957; Rogers 1962), postulates that the acceptance of some change, or the adoption of some innovation, within a population typically occurs according to a normal distribution that can be divided into the following rough groups in order of adoption: enthusiasts (innovators), visionaries (early adopters), pragmatists (early majority), conservatives (late majority), and skeptics (laggards). Here I adapt the use of these terms to group argumentation theorists according to their relative acceptance of visual arguments.

See “Appendix 1” for a proof that there are visual arguments.

Birdsell and Groarke (1996: 9) identified a series of tasks that must be met by any theory of visual argument:

any account of visual argument must identify how we can (a) identify the internal elements of a visual image (b) understand the contexts in which images are interpreted (c) establish the consistency of an interpretation of the visual, and (d) chart changes in visual perspectives over time.

Importantly, all of these are interpretative tasks relating to the identification and analysis of images, rather than anything related to their evaluation as arguments. Nor were any such items added in their revised agenda (Birdsell and Groarke 2007). It would seem, then, that Birdsell and Groarke found the evaluative apparatus to be already in place.

Additionally, Blair (2015) agrees (at least in part) with my reading of Groarke as a normative non-revisionist. He (2015: 219) writes:

Groarke, in (1996: 114), argued that visual arguments can be held to the same standards as verbal arguments. At the time, … his point was that there is no need for special dispensation for visual arguments—that no different standards are needed to admit visual arguments into the fold. … In today’s light, Groarke’s point can be seen as asserting that the probative standards (or criteria) that apply to verbal arguments may be applied equally to visual arguments, but not as denying that there might also be other probative standards that are unique to visual arguments.

In the last phrase, Blair leaves open the possibility of a normative revisionism in Groarke’s position, saying that Groarke’s position does not commit Groarke to denying the possibility of probative standards unique to visual arguments.

The source (Johnson 2010) is an unpublished conference paper that the author has graciously made available to me and permitted me to quote from.

Gilbert’s (1994, 1997) work on multi-modal argumentation does not take a position on whether the visual counts as a mode of argument. Recent correspondence confirms that, while multi-modal argumentation is intended to be open to such a possibility, Gilbert has not taken a position on this question.

See “Appendix 2” for a brief discussion of the nature of argument modality.

In identifying these as modes, Gilbert applies the concept of modality differently from the way it is applied in media, communication, and visual studies, where “mode” denotes a form of expression, and where, e.g., emotional and intuitive would not be considered modes.

The question of whether the same argument can be presented in different modes, say visual and verbal, is discussed later in Sects. 5 and 8.

Without knowing whether there actually are any self-avowed normative revisionists, this section will be speculative, seeking to articulate a set of folk assumptions and intuitions that might incline one towards a revisionist view, rather than to descriptively characterize anyone’s position. More generally, it is not relevant to the argument of the paper whether anyone actually subscribes to normative revisionism, the autonomy thesis, or the normative independence of the visual. Rather, what is important is that they are positions, with some apparent plausibility, within the spectrum of views one might take on the question of the relation between visual and non-visual (e.g., verbal) argumentation.

I prefer the view that arguments are linguistic, rather than cognitive, artifacts (Godden 2015b). Such a view needn’t commit one to the textuality of language, but it does provide what seems to me to be adequate resources for explaining the meaning, including the argumentative meaning, of images and words alike (see Groarke 2014).

This last point is important, since there is no reason to assume that we are especially reliable introspective reporters on the kinds of cognitive processes (broadly understood) that occur when we process visual information. For example, although we are typically particularly able to recognize and distinguish human faces, as well as facial expressions, our ability to do so is not a good reason for thinking that we have any good insight into, or explanation for, how we do so. I may be able to reliably make such distinctions without being able to report (whether reliably or at all) on, for example, which visual features I am processing and tracking in making these distinctions reliably. Yet, it does not follow from my inability to report on this that I am not processing, tracking, and distinguishing particular features which are articulable even if not by me.

Importantly, neither the syntactic nor semantic conditions exclude non-linguistic information. Floridi (2010: 20–21) writes:

‘[W]ell formed’ means that the data are rightly put together, according to the rules (syntax) that govern the chosen system, code, or language being used. Syntax here must be understood broadly, not just linguistically, as what determines the form, construction, composition, or structuring of something. … ‘Meaningful’ means that the data must comply with the meanings (semantics) of the chosen system, code, or language in question. Once again, semantic information is not necessarily linguistic.

Floridi offers an example of illustrations in the instruction manual for a car having a “pictorial syntax” and semantics (Floridi 2010: 21).

Note: The contention of the paper is that information properly explains the nature and content of visual artifacts (e.g., images), understood as objects of visual experience—not the nature or content of the experiences of visual objects. The paper does not contend that information (or being informed) properly characterizes the rhetorical effect or phenomenological, lived quality of experience of the object-as-experienced.

This is not to conflate the difference between saying (which collections of words do when uttered or written) and showing, or depicting, (which images do when presented or displayed). Rather, than resting on an identification of two presentational modalities, saying and showing, it relies on an identification of what is said with what is shown—i.e., of content.