The writing system used for the Native American language Cherokee (ᏣᎳᎩ ᎦᏬᏂᎯᏍᏗ Tsalagi Gawonihisdi) was invented by a Cherokee man called Sequoyah (ᏍᏏᏉᏯ Ssiquoya), a silversmith by profession, between c.1809-1821. It’s a syllabary of 85 signs, each standing for a syllable consisting of a vowel or a consonant plus a vowel, which was rapidly adopted by the Cherokee after its invention and is still in use today (the Cherokee language is endangered, with only c.2,000 first-language speakers, but the Cherokee Nation is working to promote the learning of both the language and the syllabary). The story of how Sequoyah invented this writing system is a fascinating one in its own right, but also surprisingly relevant to the study of the Bronze Age Greek syllabary Linear B – read on to find out why!
According to the most common version of the story of Sequoyah’s invention of this writing system, he was aware of English alphabetic writing, but was not himself literate (and may not even have spoken English); he therefore took some shapes of English letters to create Cherokee characters, but – after a long period of experimentation to find the best way of representing his language in writing – developed this syllabary, in which the signs which resemble English letters bear no relation to their sound-values (e.g. ᏣᎳᎩ looks like C-W-Y but spells out Tsa-la-gi ‘Cherokee’). However, Ellen Cushman, a member of the Cherokee Nation and an expert on the syllabary, points out that Sequoyah did, at least, know how to write his name in English – he had asked an acquaintance to write it out for him so that he could sign his silver work – and two English notes, written either by him or on his behalf, also exist. Moreover, the original handwritten version of Sequoyah’s syllabary has no relationship to English letters – the document below shows each character twice, on the left with its original form and on the right in the version which does show more resemblance to many English letters: this developed as a shorthand, was then used in creating the Cherokee typeface for printing, and became the main version used. Whatever Sequoyah’s precise degree of English literacy, his original creation of this writing system was thus deliberately independent of the English one; it was, and remains, an expression of the importance of language to Cherokee identity.
Relatively recent inventions of writing systems are of interest to scholars of ancient writing systems in general because for the latter, we don’t have records of their original inventor’s/inventors’ identity, motivations, and actions; historically documented instances of writing system creation provide useful parallels for thinking through reasons why people in ancient history/prehistory might have created a writing system, or might have done so in a particular way. Cherokee is particularly interesting in this respect because, unlike many writing systems created to represent the languages of peoples whose lands Europeans and settler Americans colonised, which were often the work of missionaries, it was created by a native speaker of the language in question. The way that this writing system functions in representing the Cherokee language therefore reflects the intuition of a native speaker as to the best way to do this.
I became interested in Cherokee a few years ago when I first started looking at how Linear B spelling works. Their basic structure is similar: Linear B is also a syllabary whose signs stand either for a vowel or a consonant plus vowel (occasionally two consonants).
Linear B was, of course, invented in a totally different sociocultural context from Cherokee, and via a very different process: rather than being an entirely new creation, Linear B was adapted from an existing writing system that we call Linear A. The latter was used, mainly on Crete, to write documents (mostly administrative texts on clay tablets) in a language we call Minoan (we know very little about it as we cannot fully read Linear A); the creator(s) of Linear B used many of the same signs, with probably similar sound-values but with some changes and additions, to write similar kinds of administrative texts in a different language, Mycenaean Greek. We also don’t know whether the creator(s) of Linear B was/were (a) native speaker(s) of Mycenaean Greek or of Minoan.
Given this context, there’s a tendency to assume that features of Linear B that don’t match up with the features of Mycenaean Greek, often leading to ambiguity in the way the language is written, must be due to its adaptation from a pre-existing writing system used for a totally unrelated language. For a start, the syllabic structure of Linear B, all of whose signs stand for open syllables (those ending in a vowel) does not match that of Greek, which has a large number of closed syllables (ending in a consonant) and consonant clusters (two or more consecutive consonants without a vowel between them). A syllabary like Linear B is therefore not the obvious choice of writing system to use for Greek (in contrast to, e.g., Japanese, almost all of whose syllables are open, so that representing the language in the kana syllabary of open syllables is fairly straightforward). In Linear B, consonants preceding another consonant or at the end of a word have to be either ignored in writing or written with a syllabic sign containing a ‘dummy’ vowel: the word ‘tripods’, for instance, which is tripodes in Greek, is written ti-ri-po-de, with a ‘dummy’ vowel in ti for t- and the final -s not written. Maybe Minoan was a language of (mostly) open syllables, and the creator(s) of Linear B simply retained this structure from Linear A? Moreover, Linear B is highly ambiguous in the way it represents many of the sounds of Mycenaean Greek. In the syllabary above, for instance, the signs transcribed with p– could actually stand for any of three ‘labial’ consonants (produced with the lips together): /p/ (a ‘voiceless’ labial, /ph/ (a ‘p’ with a puff of air after it, called ‘aspirated’) or /b/ (a ‘voiced labial’), and similarly for k- (/k/, /kh/, /g/) and t- (/t/, /th/ – but there is a d-series for /d/). Maybe Minoan did not have a linguistic distinction between voiceless, voiced, and aspirated consonsants – just as English, for instance, does not distinguish between voiceless and aspirated consonants (an English speaker will find it hard to hear the difference between those sounds, whereas to an ancient Greek speaker it would have been obvious) – and so had no need for separate signs to write them, and the creator(s) of Linear B kept this lack of graphic distinction despite the very different structure of their language? And finally, although most Linear B signs represent a single consonant followed by a vowel, there are a few (such as dwe, or tya) representing two consonants, though these are not consistently used (the sequence /dwe/ could be written with the sign dwe, or with two signs du-we or de-we) – could these be due to the existence in Minoan of sounds like dw and ty, leading to these signs being retained but not consistently used in Linear B?
All of these ‘problematic’ features of Linear B are, however, paralleled very closely in Cherokee. If you look at the Cherokee syllabary above, you’ll see that there are series of signs labelled g/k, d/t, and dl/tl: for some vowels, there are two separate signs for these consonants (e.g. ga, ka) but most have only a single sign standing for both consonants values (e.g. ge = ge and ke)*, leading to the exact same kind of ambiguity seen in Linear B p- etc. The Cherokee language also has consonant clusters, which are written with a ‘dummy vowel’ as in Linear B ti-ri- for /tri-/ – the word detlgv ‘trees’, for instance, is written Ꮧ Ꮱ Ꭼ de-tlu-gv – and there are a small number of signs representing clusters of two consonants, such as kw- (=kw or gw), which alternate in spelling with two-sign sequences like gv-we = gwe/kwe.
So what does this comparison with Cherokee actually tell us about Linear B? Obviously, the two languages being written are completely different, and there are other features of spelling I haven’t touched on here that work differently in the two writing systems. However, what the example of Cherokee demonstrates is that the kind of assumptions people often make about the “problems” with Linear B – that its less-than-perfect linguistic representation of Mycenaean Greek must be due to a process of adaptation, or rather lack of adaptation, from a script which itself more perfectly represented a very different language – is not necessarily true at all: some very similar features are seen in a writing system designed from scratch by a native speaker of Cherokee for the express purpose of writing that language. For Sequoyah, the advantages of having a sign for every possible syllable in the Cherokee language were clearly outweighed by the disadvantages of how large the syllabary would then become; using a single set of signs for, e.g., g- and k- and letting readers work out from context which was meant, and using ‘dummy vowels’ to represent consonant clusters, was overall a better option from the point of view of learning and using the script – while even in cases where signs like kwe were available, in practice writers often chose not to use them.
Regardless of how precisely the Linear A writing system represented the Minoan language, then – which is a whole separate question – there is no reason to think that the creator(s) of Linear B could not have made much more extensive changes to the script if they had chosen to – creating separate series of signs to distinguish all the consonants of Mycenaean Greek, adding more two-consonant signs to more easily represent consonant clusters, etc. That they did not do so is presumably due to a set of choices similar to those made by Sequoyah: precise linguistic representation was ultimately less important for the writing system’s users than the ease of learning and using a comparatively small set of signs (Cherokee’s 85 and Linear B’s 87 syllabic signs seem like a lot to those used to alphabets, but are much smaller numbers than either ancient cuneiform systems – with hundreds of signs – or, e.g., modern Chinese, with thousands. For that matter, the English writing system with its 26 letters has a far more complicated orthographic system than Linear B’s). Whenever new writing systems are created, there are always similar factors at work in balancing the size of the character set and the complexity of the orthographic conventions for the needs of the people who will use it to read and write – as well as the broader social and cultural factors that may come into play when deciding (e.g.) whether to adapt an existing writing system or create a new one. Sequoyah’s invention of the Cherokee syllabary to give his people their own means of written expression, and its continuing use by the Cherokee people today, reminds us that writing is about much more than simply representing language.
Cherokee Nation Language Department (information on learning the language and syllabary; EU-based readers may need to use a VPN to access this site)
Janine Scancarelli, ‘Aspiration and Cherokee orthographies’, in P. Downing, S.D. Lima & M. Noonan (eds), The Linguistics of Literacy (1992), p.135-152, and ‘Cherokee writing’, in P.T. Daniels & W. Bright, The World’s Writing Systems (1996), pp.587-592
W.L. Chafe & J. Kilpatrick, ‘Inconsistencies in Cherokee spelling’, in V.E. Garfield & W.L. Chafe (eds), Symposium on Language and Culture: Proceedings of the 1962 Annual Spring Meeting of the American Ethnological Society (1963), pp.60-63
On Linear B: two relevant articles of mine are available open-access:
‘Processes of script adaptation and creation in Linear B: the evidence of the “extra” signs‘ (in Philippa M. Steele (ed.), Understanding Relations Between Scripts: the Aegean Writing Systems, 2017)
‘Orthographic variation as evidence for the development of the Linear B writing system‘ (in Written Language & Literacy 22.2, 2019)
On the use of more recent writing system inventions to understand ancient ones and the processes/motivations of writing system creation in general, see work by Piers Kelly.
*for any linguists reading this, the distinction is actually between voiceless stops (transcribed, e.g., g-) and aspirated ones (transcribed, e.g., k-). Yes, this is confusing, and I assume is due to English speakers’ perception of these as sounding like voiced and voiceless due to difficulty perceiving aspiration, though I’m not sure of the exact origin of these standard transcriptions. I should also say that obviously Cherokee is not my area of study and all of my information on Cherokee linguistics is based on secondary sources like those listed above, so if any native speakers and/or researchers of Cherokee are reading this, please do let me know of any mistakes/misunderstandings/misspellings I may have committed!