I agree with the core ideas of the target article. Although there are many inscription-based sign systems, most of these, perhaps all, are specialist. A paradigm is the Chess Informator system of the 1960s; a Yugoslavian innovation that combined an existing alphanumeric convention for recording chess moves, icons based on the most common design of chess pieces, and a small set of arbitrary symbols expressing chess-specific judgements (e.g., “++” after a series of moves meant “with decisive advantage”). This innovation made it possible for the publisher in question to produce a languageless series of volume recording and analysing the main games of the preceding half year. Until the internet took over, these were the main archive of the chess world. Chess Informator shows that language-independent ideographic communication was both possible and advantageous.
This example is not unique: Music is another. As Morin shows, ideographic systems are typically specialist, designed for communicating about a specific domain. There is a gap in the space of ideographic systems, with Morin arguing convincingly that there are no general purpose, language-independent ideographic systems. Chinese scripts have been taken to be counterexamples to this claim, in part because they have been repurposed to express Japanese, and in part because a common ideogram menu was used in China despite very considerable regional differences in spoken languages. But Morin, to my mind, is convincing in arguing that these scripts are not fully language-independent. He is much less convincing in arguing that there are no general-purpose ideographic systems at all, as this obliges him to reject the view that written English, for example, is a general purpose system. He claims instead that it is a special purpose system for representing (in this case) the phonology and morphology of spoken English, saying at one point that “writing, as a code, represents language” (target article, sect. 3, para. 4) and a little later that it is, as a code “highly specialized, merely a notation of … phonemes” (target article, sect. 4, para. 1). Morin supports this with the observation that in liturgical contexts we find individuals who can read and recite sacred texts that they do not understand. Although striking, it is far from clear that this competence is productive; that they can write as well as read, or read novel texts or novel layouts of a known text.
This move misrepresents an important insight. It is true that there is a sense in which written English is parasitic on spoken (or signed) English. No one learns written English as their first or only language: Literacy is always scaffolded by a prior linguistic competence, and this is indeed relevant to explaining the gap. But this unnecessary claim pushes Morin into a strange corner. He has to deny the existence of a phenomenon once common at universities: Individuals literate in a language they cannot speak (this is the inverse of his ecclesiastic example above). He has no natural way to describe the competence of the literate but congenitally deaf, because written English has a much less direct relationship to signed English than to spoken English. Moreover, although Chinese ideograms are not language-independent, they are not plausibly described as specifications of the voicing of their spoken equivalents. Finally, it seems that Morin has to say strange things about the semantics of written sentence: If written English represents its spoken version, presumably the written sentence “Berlin is the capital of Germany” is true, because it correctly represents the structure and sounding of the spoken twin of that sentence.
The relationship between written and spoken language is asymmetrical not because written language is about spoken language, but through constraints on the acquisition of language (and to a lesser extent, its use). As Morin points out, when agents are acquiring a rich and flexible system (especially one that can be used idiosyncratically, on the fly, to cope with unexpected communicative challenges), misfires are inevitable, and efficient repair is essential. In face-to-face interaction, misfires can usually be detected and corrected quickly, because sign production is rapid, cheap, and supported by nonlinguistic signals. That allows conversation to be multiparty, free-flowing but repairable (if Steve Levinson is right, supported by cognitive mechanisms tuned to both the speed of conversational flow, and the sequential appearance of signs over time; see, e.g., Levinson, Reference Levinson and Hagoort2019). That is not true of ideographic systems. These allow individuals to communicate over space and time, but they do not support conversation at a place and time: They are not interactable in the way voice or sign is. All forms of spontaneous interaction are much more laborious. To that extent Morin is right: We cannot do everything in written English that we can do in spoken English (but also vice versa: Most of us would find it difficult to produce research papers without recourse to language in its written form). These constraints on interaction, Morin suggests, explain why literacy has to be scaffolded by voice or sign, and the same will be true of less specialised ideographic codes.
The upshot is that ideographic, language-independent general purpose systems are not impossible in principle. But their acquisition would have to be scaffolded, probably by both spoken and written forms of language. Acquiring the ability to read Informator analyses required very little investment: The sign menu is small, much of it is iconic, and sign strings are just specifications of chess moves. Acquiring a general purpose system would be very demanding: The elementary sign menu would have to be very large, iconicity would offer less help, and the combination rules would have to enable the system to approach the expressive lower of language. The up-front investment looks heavy. Why would anyone pay? A language-independent general purpose ideographic system allows users to communicate across space and time, but literacy in a shared language already allows this. Although such a system also allows communication without a shared language, the first mover burden is crippling. If a community of users already existed, it might be worth paying to join, but until then, it looks more rational to invest in learning an extra language: At least one knows, for example, that others can read French. Morin's insight into the economics of ideographic codes is independent of his eccentric analysis of literacy, and does the explanatory work he needs.
You have
Access
I agree with the core ideas of the target article. Although there are many inscription-based sign systems, most of these, perhaps all, are specialist. A paradigm is the Chess Informator system of the 1960s; a Yugoslavian innovation that combined an existing alphanumeric convention for recording chess moves, icons based on the most common design of chess pieces, and a small set of arbitrary symbols expressing chess-specific judgements (e.g., “++” after a series of moves meant “with decisive advantage”). This innovation made it possible for the publisher in question to produce a languageless series of volume recording and analysing the main games of the preceding half year. Until the internet took over, these were the main archive of the chess world. Chess Informator shows that language-independent ideographic communication was both possible and advantageous.
This example is not unique: Music is another. As Morin shows, ideographic systems are typically specialist, designed for communicating about a specific domain. There is a gap in the space of ideographic systems, with Morin arguing convincingly that there are no general purpose, language-independent ideographic systems. Chinese scripts have been taken to be counterexamples to this claim, in part because they have been repurposed to express Japanese, and in part because a common ideogram menu was used in China despite very considerable regional differences in spoken languages. But Morin, to my mind, is convincing in arguing that these scripts are not fully language-independent. He is much less convincing in arguing that there are no general-purpose ideographic systems at all, as this obliges him to reject the view that written English, for example, is a general purpose system. He claims instead that it is a special purpose system for representing (in this case) the phonology and morphology of spoken English, saying at one point that “writing, as a code, represents language” (target article, sect. 3, para. 4) and a little later that it is, as a code “highly specialized, merely a notation of … phonemes” (target article, sect. 4, para. 1). Morin supports this with the observation that in liturgical contexts we find individuals who can read and recite sacred texts that they do not understand. Although striking, it is far from clear that this competence is productive; that they can write as well as read, or read novel texts or novel layouts of a known text.
This move misrepresents an important insight. It is true that there is a sense in which written English is parasitic on spoken (or signed) English. No one learns written English as their first or only language: Literacy is always scaffolded by a prior linguistic competence, and this is indeed relevant to explaining the gap. But this unnecessary claim pushes Morin into a strange corner. He has to deny the existence of a phenomenon once common at universities: Individuals literate in a language they cannot speak (this is the inverse of his ecclesiastic example above). He has no natural way to describe the competence of the literate but congenitally deaf, because written English has a much less direct relationship to signed English than to spoken English. Moreover, although Chinese ideograms are not language-independent, they are not plausibly described as specifications of the voicing of their spoken equivalents. Finally, it seems that Morin has to say strange things about the semantics of written sentence: If written English represents its spoken version, presumably the written sentence “Berlin is the capital of Germany” is true, because it correctly represents the structure and sounding of the spoken twin of that sentence.
The relationship between written and spoken language is asymmetrical not because written language is about spoken language, but through constraints on the acquisition of language (and to a lesser extent, its use). As Morin points out, when agents are acquiring a rich and flexible system (especially one that can be used idiosyncratically, on the fly, to cope with unexpected communicative challenges), misfires are inevitable, and efficient repair is essential. In face-to-face interaction, misfires can usually be detected and corrected quickly, because sign production is rapid, cheap, and supported by nonlinguistic signals. That allows conversation to be multiparty, free-flowing but repairable (if Steve Levinson is right, supported by cognitive mechanisms tuned to both the speed of conversational flow, and the sequential appearance of signs over time; see, e.g., Levinson, Reference Levinson and Hagoort2019). That is not true of ideographic systems. These allow individuals to communicate over space and time, but they do not support conversation at a place and time: They are not interactable in the way voice or sign is. All forms of spontaneous interaction are much more laborious. To that extent Morin is right: We cannot do everything in written English that we can do in spoken English (but also vice versa: Most of us would find it difficult to produce research papers without recourse to language in its written form). These constraints on interaction, Morin suggests, explain why literacy has to be scaffolded by voice or sign, and the same will be true of less specialised ideographic codes.
The upshot is that ideographic, language-independent general purpose systems are not impossible in principle. But their acquisition would have to be scaffolded, probably by both spoken and written forms of language. Acquiring the ability to read Informator analyses required very little investment: The sign menu is small, much of it is iconic, and sign strings are just specifications of chess moves. Acquiring a general purpose system would be very demanding: The elementary sign menu would have to be very large, iconicity would offer less help, and the combination rules would have to enable the system to approach the expressive lower of language. The up-front investment looks heavy. Why would anyone pay? A language-independent general purpose ideographic system allows users to communicate across space and time, but literacy in a shared language already allows this. Although such a system also allows communication without a shared language, the first mover burden is crippling. If a community of users already existed, it might be worth paying to join, but until then, it looks more rational to invest in learning an extra language: At least one knows, for example, that others can read French. Morin's insight into the economics of ideographic codes is independent of his eccentric analysis of literacy, and does the explanatory work he needs.
Financial support
This study was supported by Australian Research Grants Council DP210102513.
Competing interest
None.