Ideography, Blissymbolics, standardization, and emergent conformity

Abstract

There is an extensive literature on the usage of Blissymbolics in augmentative and alternative communication that contradicts Morin's contention that it fails as an ideography. Morin's notion of “standardization” (target article, sect. X, para. X) is at odds with the highly developed understanding of this notion in linguistics. What Morin seems to have in mind corresponds to the notion of emergence in iterative and multiagent models of language learning.

Morin claims that Blissymbolics are unworkable:

Attempted generalist ideographies, like Bliss symbolics, struggle to express as wide as broad a variety of meaning as language does, in part because of the large number of conventional symbols that one would need to learn in order to make the system work, and in part because the rules that are supposed to help compose complex expressions from simpler symbols are too ambiguous. (target article, sect. 5.3, para. 4)

This claim appears to rest on the “evidence” from Morin, Kelly, and Winters (2020):

At the opposite end of the spectrum, one finds graphic codes like Bliss symbols (Okrent, 2010), which were invented by Charles Bliss in an attempt to create an international writing system that would allow communication between different linguistic communities. The system consists of several hundred symbols for basic concepts that can be combined to create novel or more complex expressions. Although, in principle, Bliss symbols were designed to be compositional, in practice, their users would have needed to learn as many 900 individual ideographs, along with their conventional rules of combination. No one (not even the system's inventor) really became fluent in their use.

However, this single observation is contradicted in the literature on augmentative and alternative communication. Even a cursory search through Google Scholar reveals positive results from the following reports of attempts to teach Bliss symbols, or the more recent coinage, blissymbols: Beck and Fritz (1998); Burroughs, Albritton, Eaton, and Montague (1990); Clark (1981); Ecklund and Reichle (1987); Funnell and Allport (1989); Hurlbut et al. (1982); Mizuko (1987); and Poupart, Trudeau, and Sutton (2013). Muter (1986) conducted a review with many more instances of positive results. And this is surely an undercount, because many abstracts do not indicate the name of the ideography that is used in the study. Thus it seems premature to dismiss Blissymbolics at the current stage of research.

Morin refers to the back-and-forth refinement of reference of a particular ideogram as “standardization.” This is not the accepted usage of the term in linguistics.

To refer to a readily available overview, Standard Language (2023) cites (Van Mol, 2003): “A standard variety can be conceptualized in two ways: (i) as the sociolect of a given socio-economic stratum or (ii) as the normative codification of a dialect, an idealized abstraction.”

A popular model of the latter was set forth in Haugen (1966a, 1966b), in which standardization advances in four phases. The first is norm selection, a process of choosing between competing varieties. Once a variety is settled on, it is codified by elaborating a range of reference works, such as dictionaries, grammars, spelling manuals, and style guides. This new code is “implemented” when a society accepts and transmits among speakers. Finally, the norm can be elaborated further to extend it to new functions or changing sociopolitical conditions.

In Haugen's model, Blissymbolics is most certainly standardized. Blissymbols (2023) explains that Blissymbolics Communication International published a reference guide (Wood, Storr, & Reich, 1992), containing 2,300 vocabulary items and detailed rules for the graphic design of additional characters. Thus Morin's terminology leads to considerable confusion, if not contradiction, with the facts about an existing ideography.

To be charitable, we can imagine that Morin would maintain that the implementational phase of Haugen's standardization process has never been achieved for Blissymbolics. We concede that the studies cited above are all short term and performed on small subject populations. This is why we maintain that the purported unviability of all ideographies is still up for empirical confirmation.

To our way of thinking, what Morin seems to envision as “standardization” is a kind of bottom-up, “emergent conformity” (our coinage) that arises in iterative and multiagent models of language learning, see Kirby, Griffiths, and Smith (2014) for a recent review. The literature in this paradigm tends to use the term “emergence” to label how the final state of the algorithmic process results from its initial state. In our reading, this notion of emergence is not discussed on its own merits, because it is held to be something that dynamical systems just do. What is clear, however, is that the algorithmic process is NOT labeled “standardization.”

If this terminological impasse can be resolved, we believe that Morin's analysis enters into a fruitful conversation both with iterative and multiagent modeling and with linguistic standardization. The reason is that the obstacles that Morin posits that an ideography would have to overcome to obtain wide currency (cheapness, transiency, common ground, repair, interactive alignment, lock-in, frequency-dependent advantage) are presupposed in the algorithmic architecture and deserve to be disentangled and examined on their own. In a similar vein, these factors can also be wielded to decompose Haugen's implementational phase of standardization, which, much like in the computational modeling literature, is treated as an unanalyzed black box.