[Journal of the Simplified Spelling Society, J31, 2002/2, pp31-34]

Spelling Systems Have Always Been Mixes and Have Drawn Ideas from Multiple Sources.

Niklaus Shaeffer, Basel, Switzerland.

This article is a part of a long essay covering most of the world's writing systems.
Only the sections that pertain to English have been excerpted here.

1. Introduction.

Spelling systems are heterogeneous systems that derive from several sources. There are probably no ex nihilo writing systems. Even the "Ur-Alphabet", namely the Phoenician, developed from several sources. Although there are scholars who derive alphabets such as the Runic or the Ogham alphabet from a single source, this seems very unlikely.

That English spelling is eclectic is not the problem - the problem is that English spelling mixes different systems without being consistent. Other systems - from Ancient Greek to modern Wolof spelling - are eclectic and consistent at the same time.

The creators of spelling systems are, as Miller (68) pointed out, multi-lingual and familiar with several older systems. Politics also play an important role in adapting alphabets. Those in power are able to choose their kind of alphabet. Reasons for adapting a certain alphabet may vary over time. Religion, nationalism and identity in general is probably the main force. The need to innovate and to mix different scripts is not only the result of the phonetic shape of a given language, but also due to the pressure on a political entity to have a script of their own, in order not to get confused with other groups. This political pressure also is responsible for the conservativism often encountered when it comes to changing already established systems.

English spelling with its inconsistent, historical and etymological (sometimes even pseudo-etymological) mix of the Anglo-Saxon, the Anglo-Norman and other traditions is perceived as a national symbol by many speakers of English. That English spelling is eclectic is not the problem - the problem is that English spelling mixes different systems without being consistent. Other systems - from Ancient Greek to modern Wolof spelling - are eclectic and consistent at the same time.

Nationalism has also led to some innovation. Webster's minor simplifications were generally accepted and adopted by the U.S. government. These were enough to make American spelling slightly different from British spelling. But in general English orthography since the 18th century has remained a very conservative system where reform is virtually absent.

2. The Semitic Alphabets and their Origins.

The first partly alphabetic spellings can be found in the Middle Kingdom (Sass 26). According to Bauer (Coulmas 1998: 141) the Semites borrowed the principle of consonantal alphabetic orthography (Skoyles) from the Egyptians. Gardiner (1916, in Coulmas 1989: 140) Cretan and Hittite origins are sometimes suggested. The most probable case, however, is an extensive Egyptian influence and at least graphic influence from other sources.

In the beginning, the Semitic alphabets did not contain vowel graphemes. That is why some scholars see these alphabets as syllabaries (Gelb 147 ff., Powell 238 ff.). However, syllabaries consist of items that always designate a "consonant + vowel" or in some cases "vowel". Consonantal alphabets, on the other hand, only have signs that designate consonants. It is not clear why the Semites in the beginning did not designate vowels - some scholars claim that this has something to do with the paucity of vowels present in early Semitic (as in Classical Arabic), others state that the system of Semitic roots is the cause of this system: Daniels (DB 27) claims that "the Semitic abjads do fit the structure of Hebrew, Aramaic, and Arabic very well". A syllabic system would not be fit for Semitic languages - due to the phonological properties of those languages. In Japanese, on the other hand, a consonant is always followed by a vowel, therefore syllabic writing is in a way more than fit for Japanese. The English word hotel for example is written as ho-te-ru and is also pronounced trisyllabically. Greek on the other hand is hard to write syllabically. ******v for example would have to be written as *su-ki-ze-nu. Hence, the syllabic Linear B which was used by the Mycenaeans and was derived from the earlier Minoan system (Linear A) was a system that was not created for Greek, but for a still unknown language with a phonetic structure probably similar to Japanese. Miller (18 ff.) gives the following example: /p-hásgana/ was written as pa-ka-na (id. 19) in Linear B.

2.1. Matres lectionis.

consonants used for vowel transcription
According to Sass (5), already in the Middle Kingdom there were some cases of matres lectionis, i.e. consonant graphemes which were used to transcribe vowels in foreign words, namely in Punic (Jensen 290, Naveh 62), Aramaic and Hebrew (hê, wâw, jôd; sometimes even 'šlep; Naveh 62). - In Modern Hebrew (Ivrit), this system is used in non-Biblical words such as xatûl ('cat'), which is spelled xtvl. Naveh (ibid.) notes that the earliest Aramaic and Hebrew documents already used matres lectionis. Some scholars argue that therefore the Greeks must have borrowed their alphabet from the Arameans. But the practice has older roots: the Semitic cuneiform alphabet of Ugarit (13th ct. BC) already has matres lectionis (Naveh 138).

3. The Greek Alphabet.

The fact that the Greek alphabet derives from an earlier Semitic script is uncontested, the exact source(s) of the Greek alphabet are however controversial. Sass (94) mentions the Proto-Canaanite and the Phoenician scripts, Coulmas (1989: 142) and Naveh (1979: 55) mention only the Phoenician alphabet.

The Greeks were the first people who generalized the alphabetic designation of vowels. They probably did so unconsciously, but opinions on this topic are characterized by great dissension. It is not clear whether Greeks and Semites made the distinction between vowels and consonants the same way present western civilization does. Bernal (128) mentions Phoenician colonization of Greece: bilingualism was probably rather frequent - also for economic reasons. Without communication, there is no commerce. This in turn seems to imply that - as in the case of other alphabets - primarily bilingual or multilingual people are those who adapt alphabets. The Greek alphabet is probably not the result of a unique and isolated adoption, but a multi-layered process based on several Semitic alphabets. Other scripts - Miller (52) even mentions graphic Linear B influence - may also have played a part.

Maybe it is also necessary to examine whether Greeks and Phoenicians made the same distinctions between different Semitic languages that scholars make today.

4. The Etruscan Alphabet.

The Etruscan alphabet derives from the Greek. It is however not clear whether the process of adaptation took place in Italy or in Greece/Asia Minor. It was in any case a Western Greek alphabet.

5. The Latin Alphabet.

The Latin alphabet derives mainly from the Etruscan script. According to Hammarstrom (in Jensen 521), the letters for B, D, O, X hail from a Southern Italian Greek alphabet. However, there are Etruscan abecedaria with B, D, O, X (Sampson 108). Rix (203) claims that the sound values of those letters in Latin is to be attributed to Greek influence. The letters themselves were probably all present when the Romans took over the alphabet from the Etruscans (Wachter 33).

It is uncontested that the alphabet is mainly of Etruscan origin. The sound value of C proves that clearly. Etruscan had no voiced plosives, so this symbol - derived from the Greek gamma - came to stand for the unvoiced /k/ in Etruscan - as later in Latin. Jensen (521) notes that the letters C, K, Q were originally used in Latin according to Etruscan usage: C in front of /e, i/; K in front of /a/; Q in front of /u, o/. The letters thus stand for different allophones of /k/ (in the case of Latin, also /g/ and probably the phonemes / k_w/ and /g_w/ in the case of QU and GU). These spelling rules are due to the names of the letters: gamma or gemma; kappa; qoppa or quppa (Wachter 15). In Etruscan there was no /o/, so Q was used both in front of /o/ and /u/ in Latin. Y and Z were later additions taken from the Greek alphabet. G was created approximately in the 3rd century BC by Spurius Carvilius Ruga as a modification of C (Sampson 109). * (digamma) stood for /w/ in both Etruscan and Latin, but the Romans simplified the *H-/f/combination to F /f/. The semi-vowels /w, j/ and the vowels /u, u:, i, i:/ were written with the same letters, namely V and I respectively.

6. The Runes.

The runes were created by speakers of Germanic dialects in order to write their languages. Although some scholars claim the runes to be entirely of Greek (Morris in Odenstedt 359) or Latin (Odenstedt 362) origin, most scholars view this alphabet as a script of mixed origin. Seebold (441), Krause (38 ff.), Jensen (571) and Coulmas (1996: 444 ff.) think that the Runic alphabet is a mixture of North Italic/Alpine alphabets with additional Latin influence. This most frequent school of thought is certainly more realistic than the monogenetic explanations provided by Morris and Odenstedt.

Some letters are obviously Latin in origin, for example the runes for /f/ and /r/, others remind clearly - at least on a formal level - of Alpine letters, for example the /h/-rune. There are also symbols that could be either Latin or Alpine, e.g. the /i/-rune. Bernal (36) thinks that there was also some substrate alphabet involved, Miller (62) claims that the origins of the runic alphabet are archaic-Mediterranean. Both do not specify their ideas. Miller (ibid.) also writes that the phonetic parameters on which the runic alphabet is based are ultimately clearly Semitic and links them to the scripts of Byblos and Ugarit as well as the Phoenician alphabet.

Several different Runic scripts developed, including an Anglo-Saxon system that even had different symbols for /k/ and /c/ (modern English /tS/). The latter was symbolized by the old /k/-rune; a new symbol was created for Anglo-Saxon /k/.

7. Old English - Anglo-Saxon.

and the influence of Latin Orthography.
Many languages all over the world are spelled with Latin characters. Old English, too, came to be written by means of Latin characters instead of the former Runic system. However, not all phonemes of OE had Latin counterparts. At first, /w/ and /th/ were represented by means of runic letters (wynn and thorn respectively) and eth. The Anglo-Saxons learned the Latin alphabet from Irish monks. Early manuscripts symbolized the phoneme /ð/ with the Runic thorn as well as the newly created eth. They symbolized both the voiced and the voiceless allophones of /ð/. /w/ was first symbolized by <u, uu> and then replaced by the Runic wynn (Weimann 59). Not only Irish scribes had an influence on Anglo-Saxon spelling: Greek members of the Canterbury mission introduced <y> with its Greek value /y/ (formerly /u/) into English spelling (Kniesza 26). The relationship between Northumbrian monasteries and those of northern Ireland was very close. Therefore, northern spelling conventions were closer to these sources than to the south, where wynn, thorn and eth were used. Instead, Northumbrian spelling has <th> and <u>. <th> "has always been recognized as an alternative to thorn by English writers" and "was used to transcribe Greek theta" (Scragg 2). Although Northumbrian scribes started to use thorn and wynn, they disappeared in both Scots and English - probably due to their inavailability to printers. The Lindisfarne texts have both wynn and <w> (Kniesza 29). Even before the Norman influx, Northumbrians used <ch> for /ts/. Furthermore, diacritics or doubling of vowel letters are typical of Northumbrian spelling (30). In the Middle English period, Norman influence formed and changed English spelling. <ch>, <qu> were typical French graphemes that were rare before. The simplification <sch> to <sh> (Scragg 46) and other digraphs with <-h> as their second element may have been coined after <ch> (id. 30).

In Scots, <hw> or <wh> was written as <quh> more often, and <sh> was also rare; the scribes preferring the older <sch> grapheme (Kniesza 32). Digraphs with <i> were also typical of Scots (id. 40) In this context, it is, as Kniesza (33) states, also essential to have a look round and make a similar analysis not only of the neighbouring dialects (of Scots] but of all languages whose speakers played an important historical-cultural role in the life of Scotland: French, Dutch/Flemish and soon. This is because writing can be influenced externally, and the adoption of a certain spelling habit does not necessarily mean the adoption of the underlying pronunciation, let alone a parallel process in sound changes.

Scragg (17) has looked more closely at "foreign" influences. First, Latin conventions lead to a certain degree of confusion: <th> and <ch> were introduced and <ae> was often used instead of <e>. The sounds [ç] and [x] were sometimes represented by Anglo-Norman <s> as well as Old English [h] and the new grapheme yogh and yogh plus <h> (23) - probably due to the absence of these sounds in French. Scragg (49) mentions the introduction of the new <ie> grapheme which was taken over from French. Even Spanish influence may be present in English spelling, as Scragg (57) notes, namely the <l> in the word emerald which is sometimes ascribed to sixteenth century Spanish influence.

In 1476, William Caxton established the first press in England (Scragg 5). Caxton also translated himself, and according to Scragg (66) "he seems heavily influenced by his sources, the most notorious of his permanent contributions to the language being the introduction of the Dutch convention <gh> for /g/ in ghost, a native word spelt gost until the later fifteenth century." <gh> may actually be a Dutch grapheme introduced into English; however, the grapheme was almost certainly not pronounced as /g/.

The spelling in Middle Dutch was in any case strongly influenced by French writers. In old texts we can, for example, often find -ghe- or -ghi- instead of the modern -ge- or -gi-. This means that the letter <g> must have been pronounced [x] as in modern Dutch (ie like the <ch> in Scottish "loch"), and not like the French <g> which had changed to voiced <zj> in front of /e/ and /i/ (as in the English word "leisure"). (Dünser et al.) English translation by Dr. John Gledhill.

Therefore, Caxton's choice of spelling is probably solely graphic and not based on phonetic properties. He also used other "Dutch spellings". In one of his translations from the Dutch, Reynard the Fox (1481), he wrote goed instead of good and ruymen for make room. Dutch goed was at this time probably already pronounced as /u/. Words like good, foot, stood had alternative pronunciations with /u:/ until the 17th century (Cruttenden 113).

The <uy> spelling in ruymen seems to imply a diphthongized pronunciation (or maybe an older Dutch pronunciation /u:/).

The phoneme /dʒ/ that was symbolized in OE either as <cg> or in some cases <gg> due to contact with Old French orthography was now spelled either as <j> or <g> (only before <e, i>). (66/67) The spelling <gu-> before <i, e> was also due to French influence: guest instead of gest. 16th century Italian influence lead to spellings such as ghest or ghess which reflect the Italian way of distinguishing between palatal and non-palatal pronunciations before <i, e> (ibid.)

Both Scots and English spelling - as we have seen, Scots and English were probably more heterogeneous at an earlier stage which makes the term "Anglo-Saxon" more fit to describe the language - have been subject to many different influences.
1. Irish
2. Latin
3. French
4. Dutch
5. Italian
6. Spanish
And many others - e.g. Icelandic (geysir), Gaelic (loch), Portuguese (piranha), German and German transcription of Yiddish (dachshund, schmuck). English and Scots spelling is a living example of a system that has various origins and even applies different rules at the same time, due to the different systems that formed English spelling in the past.

8. Conclusion.

Monogenetic theories are usually unfit to explain the origins of alphabets and writing systems. This is true in both older alphabets as well as orthography of modern languages. There have been probably no ex nihilo alphabets or writing systems. Alphabets that have been created from a single source are very hard to find. As culture in general is always eclectic, so are alphabets and other scripts. Most often multilingual people who already know several tongues and scripts create new scripts. Modern day spelling systems are no exception to this rule. The heterogeneous origins of spelling systems are also the result of politics. Especially in modern times, nations identified themselves not only with their language, but also with their script.

Partial list of references.

Bernal, Martin. 1990. Cadmean Letters. Winona Lake: Eisenbrauns.

Berry, Jack. 1977. "The 'Making of Alphabets' Revisited" In: Fishman 3-16.

Coulmas, Florian. 1989. The Writing Systems of the World. Oxford: Basil Blackwell Ltd.

-. 1985. Sprache und Staat. Studien zur Sprachplanung. Berlin and NY: Walter de Gruyter.

-. 1996. The Blackwell encyclopaedia of writing systems. Oxford: Blackwell Publishers Ltd.

Crystal, David. 1997. The Cambridge Encyclopedia of the English Language. Cambridge: CUP.

Daniels, Peter T. and William Bright. 1996. The World's Writing Systems. NY and Oxford: OUP 1996. (=DB)

Diringer, David. 1986. The Alphabet. A Key to the History of Mankind. 3rd. Ed. Vols. I and II. Completely revised. London: Hutchinson.

Driver, G. R. 1976. Semitic Writing. From Pictograph to Alphabet. Newly revised edition. Oxford: OUP.

Dünser, Julian et al. 1997. Geschiedenis van het Nederlands

Elliott, Ralph W. 'The runic script" In: Daniels/Bright S. 333-339.

Everson, Michael et al. 1997. "On the status of the Latin letter þorn and of its sorting order"

Firmage, Richard A. 1993. The alphabet abecedarium. Some notes on letters. Boston: David R. Godine, Publisher.

Gelb, I.J. 1969 11952] A Study of Writing. Chicago and London: The U of Chicago P.

Gordon, A.E. 1973. The Letter Names of the Latin alphabet. Berkley, LA and London: U of California P.

Gundersen, Dag. 1977. "Successes and Failures in the Reformation of Norwegian Orthography" In: Fishman 247-265.

Harris-Northall, Ray. 1992. "Devoicing, deaffrication and word-final - z" In: Hispanic Linguistics 4:2 (Spring 1992)

Hooker, J.T (introd.) 1990. Reading the Past. Ancient Writing from Cuneiform to the Alphabet. London: British Museum Press.

Izady, Mehrdad R. "The Concise Encyclopaedia of Kurdistan. A Pan-Kurdish Alphabet."

Jensen, Hans. 1970. Sign Symbol and Script. London: George Allen and Unwin Ltd. Transl. of Die Schrift in Vergangenheit and Gegenwart. VEB Deutscher Verlag der Wissenschaften. 1958, as revised by the author.

Miller, D. Gary. 1994. Ancient scripts and phonological knowledge. Amsterdam/Philadelphia: John Benjamins Publishing Company.

Powell, Barry B. 1991. Homer and the Origin of the Greek Alphabet. Cambridge: CUP.

Sampson, Geoffrey. 1985. Writing Systems. London (etc.): Hutchinson.

Sass, Benjamin. 1991. Studia Alphabetica. On the origin and early history of the Northwest Semitic, South Semitic and Greek alphabets. CH-Freiburg: Universitätsverlag Freiburg Schweiz. Göttingen: Vandenhoeck & Ruprecht.

Scragg, D.G. 1974. A History of English Spelling. MUP/Barnes & Noble.

Skoyles, John R. 1977. The origin of Classical Greek culture: Hunter-gatherers of the alphabet.

Swiggers, Pierre. 1996. "Transmission of the Phoenician script to the west" In: Daniels/Bright S.261-270.

Threatte, Leslie. 1996. "The Greek alphabet" In: Daniels/Bright S. 271-280.

Back to the top.