Debates over the nature and feasibility of ideography are not going away anytime soon. Often, these have not advanced us much further than Edgerton (Reference Edgerton1941), who noted over 80 years ago the widespread presence of ideograms in English (and indeed many other writing systems), such as numerals and signs like $, which cannot be subsumed under terms like “logogram.” Morin has undertaken a salutary task here – asking not only whether ideograms exist (they surely do), nor even whether fully ideographic notations can exist, but how, where, and why they exist – and conversely, why domain-general, self-sufficient notations are rare or nonexistent. By drawing our attention to how notations are shared and transmitted, Morin has laid a pathway for future empirical studies both in cognitive science and in the field of writing systems.
We should be careful not to conflate Liberman's “openness” or Morin's “generality” with the number of messages capable of being conveyed. A notation can be infinite while still being highly restricted. Many numerical notations are infinite (adding another zero to the end, or another tally-mark to the score) and musical notations are capable of encoding an infinite number of musical themes and scores. It is that they are domain-specific, rather than that they are infinite, that should attract our attention. But this raises the question: What affords a notation this infinity? For Chomsky (Reference Chomsky1980), the relevant property is what he calls “discrete infinity” – that a limited set of discrete elements can be combined to produce a potential infinity of representations of the sort necessary for language. But this is precisely what is needed for nonlinguistic codes as well. A numerical notation which required that users find a different sign for every number would require an impossibly inhuman mind, and, of course, would constantly require recourse to inventing new symbols (Chrisomalis, Reference Chrisomalis2020, pp. 1–26).
So, when we are looking for notations that will work – that humans will be able to accept and share – we should be looking for systems that have a componential or composite structure (Chrisomalis, Reference Chrisomalis2018). It is easy to be misled by issues of modality (visual vs. auditory) here. Sign languages are for all relevant purposes, languages, and have all the properties that languages have (including phonology – see Brentari, Reference Brentari2019) and are scaffolded by the same cognitive capacities that underlie spoken languages. A general-purpose system of nonlinguistic gestures would be just as nonfunctional as an emoji-only pseudo-language. As Morin points out in the case of Blissymbolics, composite structure is not a sufficient condition, but I believe it is a necessary one. We need more inquiry into how much or what kinds of componential structure make a notation shareable and learnable.
The problem with ideographies is that many ideas are not easily encoded componentially – or at least, no one has done so without also encoding a lot of linguistic information. Given that languages are universal to humans, and languages already have a componential structure, perhaps this is not so surprising. Why bother with a new system for ideas when every speaker already has one at hand? Parts of systems – such as Chinese radicals or Egyptian hieroglyphic determiners – can be described as ideographic or semasiographic, but these are far less important to these scripts than the myths tell (DeFrancis, Reference DeFrancis1986). Perhaps we have imagined that ideographies are far more desirable than they really are. The obsession of Leibniz, Wallis, and other early modern Europeans with universal language schemes (and their concomitant interest in Chinese writing) may be a solution to a problem no actual script inventor actually had (Knowlson, Reference Knowlson1975).
Morin turns our attention, rightly, to “lock-in effects” that facilitate standardization and consistency in graphic notations as opposed to language, which develops differently because of its synchronous and face-to-face qualities. But of course notations do change, and sometimes quite rapidly. The transformation of the West African Bamum script from a logographic or ideographic notation to an alphasyllabic one took place within a generation (Kelly, Reference Kelly, Ferrara and Valério2018) under rapidly changing conditions of colonization and resistance. At a somewhat longer scale, the introduction of phonography as proto-cuneiform transformed into Sumerian and later scripts (while retaining important nonphonographic elements) shows how an essentially translinguistic notation need not be eternally so. Although the strongest teleological, unilineal evolutionary arguments are clearly wrong, there are nonetheless important cultural-evolutionary patterns on display (Trigger, Reference Trigger1998). If Morin is right, we should be able to find evidence that standardization and social sharing motivated these diachronic transformations.
Moreover, there are highly unusual systems that deserve more attention, precisely because their combinatorial semiotics defy simple labels such as “phonography” and “ideography.” For instance, the symbols of the Western Apache shaman Silas John (Basso & Anderson, Reference Basso and Anderson1973) encoded semantic, kinesic, and pragmatic information in addition to recording the words associated with particular prayers. Some of the challenges faced in analyzing systems like the notations of the Iron Age Scottish Picts (Lee, Jonathan, & Ziman, Reference Lee, Jonathan and Ziman2010) or the rongorongo of Rapa Nui (Valério & Ferrara, Reference Valério and Ferrara2019) may rest in relying on a too-narrow binary, where every system must be a pure record of written language, or else it loses theoretical interest. We should abandon this byproduct of a progressivist mindset in which writing is the gold standard against which other notations are evaluated.
Finally, although Morin does not make much of the distinction between “ideography” and “semasiography,” his interest in standardization and sharing motivates exactly this distinction. The terminological issue may seem trivial or driven by the aim of creating new terms for their own sake, but it is neither. Ideography focuses on the idea, and thus on individual cognition, whereas semasiography draws our attention to the sema, the sign, and thus to how meanings are shared among interlocutors. The semiotic focus on what makes the interindividual sharing and transmission of signs and their meanings difficult, rather than what makes it conceptually hard for any individual to grasp a notation, suggests that semasiography is really what we ought to be talking about here.
You have
Access
Debates over the nature and feasibility of ideography are not going away anytime soon. Often, these have not advanced us much further than Edgerton (Reference Edgerton1941), who noted over 80 years ago the widespread presence of ideograms in English (and indeed many other writing systems), such as numerals and signs like $, which cannot be subsumed under terms like “logogram.” Morin has undertaken a salutary task here – asking not only whether ideograms exist (they surely do), nor even whether fully ideographic notations can exist, but how, where, and why they exist – and conversely, why domain-general, self-sufficient notations are rare or nonexistent. By drawing our attention to how notations are shared and transmitted, Morin has laid a pathway for future empirical studies both in cognitive science and in the field of writing systems.
We should be careful not to conflate Liberman's “openness” or Morin's “generality” with the number of messages capable of being conveyed. A notation can be infinite while still being highly restricted. Many numerical notations are infinite (adding another zero to the end, or another tally-mark to the score) and musical notations are capable of encoding an infinite number of musical themes and scores. It is that they are domain-specific, rather than that they are infinite, that should attract our attention. But this raises the question: What affords a notation this infinity? For Chomsky (Reference Chomsky1980), the relevant property is what he calls “discrete infinity” – that a limited set of discrete elements can be combined to produce a potential infinity of representations of the sort necessary for language. But this is precisely what is needed for nonlinguistic codes as well. A numerical notation which required that users find a different sign for every number would require an impossibly inhuman mind, and, of course, would constantly require recourse to inventing new symbols (Chrisomalis, Reference Chrisomalis2020, pp. 1–26).
So, when we are looking for notations that will work – that humans will be able to accept and share – we should be looking for systems that have a componential or composite structure (Chrisomalis, Reference Chrisomalis2018). It is easy to be misled by issues of modality (visual vs. auditory) here. Sign languages are for all relevant purposes, languages, and have all the properties that languages have (including phonology – see Brentari, Reference Brentari2019) and are scaffolded by the same cognitive capacities that underlie spoken languages. A general-purpose system of nonlinguistic gestures would be just as nonfunctional as an emoji-only pseudo-language. As Morin points out in the case of Blissymbolics, composite structure is not a sufficient condition, but I believe it is a necessary one. We need more inquiry into how much or what kinds of componential structure make a notation shareable and learnable.
The problem with ideographies is that many ideas are not easily encoded componentially – or at least, no one has done so without also encoding a lot of linguistic information. Given that languages are universal to humans, and languages already have a componential structure, perhaps this is not so surprising. Why bother with a new system for ideas when every speaker already has one at hand? Parts of systems – such as Chinese radicals or Egyptian hieroglyphic determiners – can be described as ideographic or semasiographic, but these are far less important to these scripts than the myths tell (DeFrancis, Reference DeFrancis1986). Perhaps we have imagined that ideographies are far more desirable than they really are. The obsession of Leibniz, Wallis, and other early modern Europeans with universal language schemes (and their concomitant interest in Chinese writing) may be a solution to a problem no actual script inventor actually had (Knowlson, Reference Knowlson1975).
Morin turns our attention, rightly, to “lock-in effects” that facilitate standardization and consistency in graphic notations as opposed to language, which develops differently because of its synchronous and face-to-face qualities. But of course notations do change, and sometimes quite rapidly. The transformation of the West African Bamum script from a logographic or ideographic notation to an alphasyllabic one took place within a generation (Kelly, Reference Kelly, Ferrara and Valério2018) under rapidly changing conditions of colonization and resistance. At a somewhat longer scale, the introduction of phonography as proto-cuneiform transformed into Sumerian and later scripts (while retaining important nonphonographic elements) shows how an essentially translinguistic notation need not be eternally so. Although the strongest teleological, unilineal evolutionary arguments are clearly wrong, there are nonetheless important cultural-evolutionary patterns on display (Trigger, Reference Trigger1998). If Morin is right, we should be able to find evidence that standardization and social sharing motivated these diachronic transformations.
Moreover, there are highly unusual systems that deserve more attention, precisely because their combinatorial semiotics defy simple labels such as “phonography” and “ideography.” For instance, the symbols of the Western Apache shaman Silas John (Basso & Anderson, Reference Basso and Anderson1973) encoded semantic, kinesic, and pragmatic information in addition to recording the words associated with particular prayers. Some of the challenges faced in analyzing systems like the notations of the Iron Age Scottish Picts (Lee, Jonathan, & Ziman, Reference Lee, Jonathan and Ziman2010) or the rongorongo of Rapa Nui (Valério & Ferrara, Reference Valério and Ferrara2019) may rest in relying on a too-narrow binary, where every system must be a pure record of written language, or else it loses theoretical interest. We should abandon this byproduct of a progressivist mindset in which writing is the gold standard against which other notations are evaluated.
Finally, although Morin does not make much of the distinction between “ideography” and “semasiography,” his interest in standardization and sharing motivates exactly this distinction. The terminological issue may seem trivial or driven by the aim of creating new terms for their own sake, but it is neither. Ideography focuses on the idea, and thus on individual cognition, whereas semasiography draws our attention to the sema, the sign, and thus to how meanings are shared among interlocutors. The semiotic focus on what makes the interindividual sharing and transmission of signs and their meanings difficult, rather than what makes it conceptually hard for any individual to grasp a notation, suggests that semasiography is really what we ought to be talking about here.
Competing interest
None.