Origin of SCARLET
Posted: 24 November 2010 04:02 PM   [ Ignore ]
RankRank
Total Posts:  64
Joined  2010-11-02

The word “scarlet” is attested in Latin about year 1100 spelled ‘’scarlata’’ and meaning expensively dyed bright cloth of various colors. When bright, rich dye was used in medieval clothing, red was the most common color. “Scarlet” as a color, which is in all European languages today, started out in late medieval times referring to scarlet-dyed clothing. Question: What was the source of the medieval Latin word? That’s a tough question and I’m not going to try to answer it. A 12-page article devoted to it is at ref (written in 1913), and further info written more recently is at ref (written in French). But I want to look at how today’s ordinary English dictionaries handle this tough question in their summary etymology of SCARLET. (This is one of a series of posts arguing that you can’t trust any of today’s ordinary English dictionaries to reliably summarize what’s known about the etymology of a word). Looking at the following seven English dictionaries, their summaries can be divided into three classes:

CLASS ONE:
Concise OED @ OxfordDictionaries.com: Middle English (originally denoting any brightly coloured cloth): shortening of Old French escarlate, from medieval Latin scarlata, via Arabic and medieval Greek from late Latin sigillatus ‘decorated with small images’, from sigillum ‘small image’
Random House @ Dictionary.com: 1200–50; ME < OF escarlate < ML scarlata, scarletum, perh. < Ar saqirlāṭ, siqillāṭ < MGk sigillátos < L sigillātus decorated with patterns in relief; see sigillate

CLASS TWO:
Merriam-Webster @ M-W.com: Middle English scarlat, scarlet, from Anglo-French escarlet, from Medieval Latin scarlata, from Persian saqalāt, a kind of rich cloth. First Known Use in English: 13th century
Webster’s New World @ YourDictionary.com: ME < OFr escarlate < ML scarlatum < Pers säqirlāt, dress dyed crimson < ?
Chambers Dict @ ChambersHarrap.co.uk: 13c: from French escarlate, from Persian saqalat scarlet cloth.
American Heritage @ YourDictionary.com: Middle English, scarlet cloth, scarlet, from Old French escarlate, from Medieval Latin scarlata, scarlet cloth, from Persian saqirlāt, rich cloth, scarlet cloth, variant of siqillāt, from Arabic, perhaps from Medieval Greek *sigillatos, from Latin sigillātus, decorated with raised figures, from sigilla, little figures, pl. of sigillum, sigil; see sigil.

CLASS THREE:
Collins English @ Dictionary.com: 13th century English: from Old French escarlate fine cloth, of unknown origin

The majority of the dictionaries are in the class that says with no uncertainty that the Latin was borrowed directly from Persian. To evaluate it, consider the following five points. (1) Latin had no direct contact with Persian in that era. Any borrowing from Persian would’ve had to have been mediated by Arabic or by Byzantine Greek. (2) I haven’t seen any citation of any use of a Persian saqirlāt or saqalāt in Persian that antedates the year 1100 attestation of Latin ‘’scarlata’’ and I must infer that no such attestation exists. (3) The Persian word ‘’saqirlāṭ’’ occurs in late 13th century Persian defined as “a woollen fabric made in the land of the Franks [i.e. Christian Europe]”. (4) The wordforms ‘’siqillat’’ and ‘’sijillat’’ are attested multiple times in Arabic in the 9th century as the name of a type of cloth. They are also attested in 10th and 11th century Arabic meaning a type of cloth. The Arabic was definitely borrowed from early medieval Greek ‘’sigillatos’’, a type of cloth, which definitely comes from late classical Latin ‘’sigillatum’’. It is extremely likely that the later-seen Persian words of the form siqillāṭ and saqallāṭ are borrowed from this Arabic. Once again, there is no attestation for those Persian words in Persian pre-1100 to the best of my knowledge. (5) The reason why the majority of the dictionaries are deriving Latin ‘’scarlata’’ from Persian ‘’saqirlāt’’ is: (a) ‘’scarlata’’ has no precedents in Latin from which it might be derived within Latin; (b) even though early Arabic has lots of ‘’siqillat’’ attestations, the form ‘’saqirlāt’’ with the ‘r’ is absent from Arabic (except at a date far too late) and the derivation for ‘’scarlata’’ needs the ‘r’ of ‘’saqirlāt’’ to get ‘’scarlata’’; and (c) the form ‘’saqirlāt’’ with the ‘r’ is attested in medieval Persian.

In conclusion, the method of the majority of the dictionaries appears to me to be: ‘’saqirlāt’’ exists and it can be fitted up to ‘’scarlata’’, let’s just do that, and to hell with all the historical-context deficiencies that it has.

[ Edited: 24 November 2010 04:32 PM by ᴚǝǝƶɐʍɐɈ ]
Profile
 
 
Posted: 24 November 2010 04:47 PM   [ Ignore ]   [ # 1 ]
RankRankRankRankRank
Total Posts:  3524
Joined  2007-01-29

Well, I’m glad you got that off your chest.  But:

(1) What era?

(2) What sources do you have for historical attestation in Persian? (Secondary observation: The argumentum ex silentio is not a reliable one.)

(4) Why is it “extremely likely” that Persian would have randomly stuck an -r- into a borrowed Arabic word?  Are you aware of any similar examples?

Profile
 
 
Posted: 24 November 2010 06:47 PM   [ Ignore ]   [ # 2 ]
RankRank
Total Posts:  64
Joined  2010-11-02

Latin had no direct contact with Persian in the era when the word ‘’scarlata’’ entered the written Latin records, that is, the medieval centuries immediately preceding the first Latin attestation in 1100.

The rest of what I reported above is all drawn from the two online sources I linked to at the beginning of the post. One thing the French one mentions is that in Persian word appears in the forms siqillāṭ, saqallāṭ, saqallāt, saqirlāṭ, saqirlāt, and saghirlāṭ. It’s worth adding that varieties of the word also appear in India, borrowed from the Persian, and some info about those is in the book by Yule & Burnell (1903) online here.

languagehat asks: “Why is it “extremely likely” that Persian would have randomly stuck an -r- into a borrowed Arabic word?” I gather I’ve been misunderstood, because I wasn’t suggesting such a thing. What I said was that the Persian of the form ‘’siqillat’’ is extremely likely to have been borrowed from the Arabic of the form ‘’siqillat’’. The reasons for the exteme likilhood are (a) the word certainly comes from early medieval Greek, and Arabs adopted lots of words directly from Greek with many of such words later entering Persian from the Arabic, whereas Persian didn’t borrow directly from Greek nearly so much as Arabic did in the medieval centuries preceding 1100, and (b) the Arabic is attested in the 9th century, and (c) the Persian ‘’siqillat’’ form is so extremely close to the Arabic ‘’siqillat’’ form than one must’ve borrowed from another, one way other the other, and (d) I haven’t heard of any dated attestation in Persian until centuries later than the 9th.

Profile
 
 
Posted: 24 November 2010 08:24 PM   [ Ignore ]   [ # 3 ]
Administrator
Avatar
RankRankRankRankRank
Total Posts:  4784
Joined  2007-01-03

Latin had no direct contact with Persian in the era when the word ‘’scarlata’’ entered the written Latin records, that is, the medieval centuries immediately preceding the first Latin attestation in 1100.

This is a rather absurd statement. If there is one thing that history teaches us, is that people get around more than you might think. There was certainly trade between Europe and Persia in those centuries. Not a lot perhaps, but to say there was none and no chance of direct linguistic contact (particularly regarding a valuable trade good like dye) is just silly.

As an illustration on this Thanksgiving holiday, I remind you that when my forefathers landed at Plymouth they were met by an Indian who spoke English and had been to England and back. People get around.

Profile
 
 
Posted: 25 November 2010 01:08 AM   [ Ignore ]   [ # 4 ]
RankRankRankRank
Total Posts:  837
Joined  2007-03-01

The overland “Silk Route” from China to the West ran right through Persia, and one of the sea routes went up the Persian Gulf. If there were any points of contact at all between Persia and Latin-writing Europe in the centuries immediately preceding 1100, the luxury textiles trade was certainly one of them.

Profile
 
 
Posted: 25 November 2010 01:37 AM   [ Ignore ]   [ # 5 ]
RankRank
Total Posts:  44
Joined  2010-09-07
ᴚǝǝƶɐʍɐɈ - 24 November 2010 04:02 PM

(b) even though early Arabic has lots of ‘’siqillat’’ attestations, the form ‘’saqirlāt’’ with the ‘r’ is absent from Arabic (except at a date far too late) and the derivation for ‘’scarlata’’ needs the ‘r’ of ‘’saqirlāt’’ to get ‘’scarlata’’

Not true.  According to Corominas (Diccionario Crítico Etimológico Castellano e Hispánico), the Mozarab (Spanish Arabic) form ’iskirlata is found in 1001.  Corominas’ path of derivation is thus

SIGILLATUM (Latin) -> sigillatos (Greek) -> siqillat (classical Arabic) -> skirlata/iskirlata (Sp. Arabic)

From Spanish Arabic (or an unattested early Spanish source) it would then have passed on to Medieval Latin.  For a word history of scarlet from a Spanish perspective, see the account in Spanish Vocabulary: an Etymological Approach (disclaimer: I am the author). Most studies that have looked into the matter in detail conclude that the Persian comes from the Arabic: Corominas, the French CNRTL, the 1913 article cited in the first post in this series, and others.

Profile
 
 
Posted: 25 November 2010 01:49 AM   [ Ignore ]   [ # 6 ]
RankRank
Total Posts:  64
Joined  2010-11-02
Syntinen Laulu - 25 November 2010 01:08 AM

The overland “Silk Route” from China to the West ran right through Persia, and one of the sea routes went up the Persian Gulf. If there were any points of contact at all between Persia and Latin-writing Europe in the centuries immediately preceding 1100, the luxury textiles trade was certainly one of them.

There was lots of trade including lots of textiles trade. But Persian wasn’t one of the languages of trade (except locally in Persia). The main language of trade was Arabic. Greek was also used as a language of trade in the eastern Mediterranean. To pass from Greek-speaking territory to Persian-speaking territory, a merchant would pass through a big territory where Arabic was the language of business, and when he arrived in Persia he’d find that Arabic was still the language of business there too.

During the centuries 700 - 1100 Arabic was the lingua franca of the Islamic countries.

Most of the writers living in Persian-speaking territory wrote in Arabic during those centuries.

If a word is found in Arabic and not in Persian, 800 - 1100, it may still be an Arabic borrowing from Persian. The well-known writer Al-Razi (died around 930) grew up in Persian territory and wrote in Arabic, but his writings have lots of Persian names for things, especially Persian botanical names. Some of Al-Razi’s names are not attested in Persian before he used them in Arabic, but they can still be labelled Persian names with high confidence because of their phonetic structure and because of their absence from more westerly Arabic (and Greek) writers. In light of the relative scarcity of Persian writings 800 - 1100, the non-appearance of ‘’siqillat’’ in Persian writings is not a great reason to believe it wasn’t in Persian at all. However, no matter what the earliest Persian ‘’siqillat’’ might be, there is a great reason to believe that it is a borrowing from the Arabic ‘’siqillat’’, and I’ve said it twice already, namely, ‘’siqillat’’ is descended from early medieval Greek ‘’sigillatos’’ which is descended from classical Latin ‘’sigilla’’. The article I linked to earlier gives other reasons to believe the Persian is from Arabic.

Profile
 
 
Posted: 25 November 2010 02:02 AM   [ Ignore ]   [ # 7 ]
RankRank
Total Posts:  64
Joined  2010-11-02
madeira - 25 November 2010 01:37 AM

ᴚǝǝƶɐʍɐɈ - 24 November 2010 04:02 PM
the form ‘’saqirlāt’’ with the ‘r’ is absent from Arabic (except at a date far too late)

Not true.  According to Corominas (Diccionario Crítico Etimológico Castellano e Hispánico), the Mozarab (Spanish Arabic) form ’iskirlata is found in 1001.... The French CNRTL says the same.

I saw that at CNRTL. But “Mozarab” is not Arabic and not Spanish Arabic. It is Spanish. Here’s the defintion Wikipedia:

Mozarabic was a continuum of closely related Romance dialects spoken in Muslim-dominated areas of the Iberian Peninsula during the early stages of the Romance languages’ development in Iberia. Mozarabic descends from Late Latin and early Romance dialects spoken in the Iberian Peninsula from the 5th to the 8th centuries, and was spoken until the 14th century.

Other than that, madeira, thanks for the link. Another high caliber post.

Profile
 
 
Posted: 25 November 2010 03:07 AM   [ Ignore ]   [ # 8 ]
RankRank
Total Posts:  44
Joined  2010-09-07
ᴚǝǝƶɐʍɐɈ - 25 November 2010 02:02 AM

madeira - 25 November 2010 01:37 AM
ᴚǝǝƶɐʍɐɈ - 24 November 2010 04:02 PM
the form ‘’saqirlāt’’ with the ‘r’ is absent from Arabic (except at a date far too late)

Not true.  According to Corominas (Diccionario Crítico Etimológico Castellano e Hispánico), the Mozarab (Spanish Arabic) form ’iskirlata is found in 1001.... The French CNRTL says the same.

I saw that at CNRTL. But “Mozarab” is not Arabic and not Spanish Arabic. It is Spanish. Here’s the defintion Wikipedia:

Mozarabic was a continuum of closely related Romance dialects spoken in Muslim-dominated areas of the Iberian Peninsula during the early stages of the Romance languages’ development in Iberia. Mozarabic descends from Late Latin and early Romance dialects spoken in the Iberian Peninsula from the 5th to the 8th centuries, and was spoken until the 14th century.

Other than that, madeira, thanks for the link. Another high caliber post.

You’re right.  But the fact that people in Arabic Spain were using iskirlata can be taken as strong evidence that it was being used in Spanish Arabic as well (where it is, according to Corominas, subsequently attested, as well as in the Arabic of Morocco).

Profile
 
 
Posted: 25 November 2010 05:47 AM   [ Ignore ]   [ # 9 ]
RankRankRankRankRank
Total Posts:  3524
Joined  2007-01-29

The rest of what I reported above is all drawn from the two online sources I linked to

That’s what I was afraid of.  You can’t do etymology (or anything else of any value) by relying on online sources.

But Persian wasn’t one of the languages of trade (except locally in Persia).

This is utterly wrong; Persian was in fact the major lingua franca everywhere between Syria and China.

Profile
 
 
Posted: 25 November 2010 06:44 AM   [ Ignore ]   [ # 10 ]
Avatar
RankRankRankRank
Total Posts:  2344
Joined  2007-01-30

Verb. sap. Old Persian proverb: he who build long detailed argument on free internet resources build house on sand.

Profile
 
 
Posted: 25 November 2010 07:17 AM   [ Ignore ]   [ # 11 ]
Administrator
Avatar
RankRankRankRankRank
Total Posts:  4784
Joined  2007-01-03

You can’t do etymology (or anything else of any value) by relying on online sources.

Generally, I agree and the point is definitely applicable to ᴚǝǝƶɐʍɐɈ’s arguments. But it should be qualified somewhat.

First, I would say “relying solely.” I would also distinguish between free resources and subscription services. There are many excellent online resources that are only available through research libraries or via paid subscription. (The OED springs immediately to mind. You can’t do serious etymological work in English without access to the OED; although the OED is not the end all and be all. You need other sources as well. Journal articles are another, which are mostly online now.)

But you can still do valuable work in certain fields without paid online services. Researching current slang is one such. There’s lots of good things that can be done without paying a dime. In another, Brigham Young University makes available modern corpuses of literature for free that can be used in all sorts of linguistic research. I’m sure other fields will have similar examples. But it is true that for the type of work that ᴚǝǝƶɐʍɐɈ is trying to do, the free online sources are only the start of the research.

Profile
 
 
Posted: 26 November 2010 03:02 AM   [ Ignore ]   [ # 12 ]
RankRank
Total Posts:  64
Joined  2010-11-02
madeira - 25 November 2010 03:07 AM

the fact that [non-Arabic] people in Arabic Spain were using iskirlata can be taken as strong evidence that it was being used in Spanish Arabic as well (where it is, according to Corominas, subsequently attested, as well as in the Arabic of Morocco).

You’ve seen my sources youself, and you know they’ve no attestation of a ‘’siqirlat’’ wordform in medieval Arabic. That is strongly suggestive evidence that the word was not being used in Arabic. With regard to what Corominas says, I take it he’s referring to relatively modern attestations in Arabic. Yule & Burnell make the caution: “The fact that the [post-medieval] Arabic dictionaries give a form ‘’sakirlat’’ must not be trusted to. It is a modern form, probably taken from the European word.”

Profile
 
 
Posted: 26 November 2010 03:36 AM   [ Ignore ]   [ # 13 ]
RankRank
Total Posts:  44
Joined  2010-09-07
ᴚǝǝƶɐʍɐɈ - 26 November 2010 03:02 AM

madeira - 25 November 2010 03:07 AM
the fact that [non-Arabic] people in Arabic Spain were using iskirlata can be taken as strong evidence that it was being used in Spanish Arabic as well (where it is, according to Corominas, subsequently attested, as well as in the Arabic of Morocco).

You’ve seen my sources youself, and you know they’ve no attestation of a ‘’siqirlat’’ wordform in medieval Arabic. That is strongly suggestive evidence that the word was not being used in Arabic. With regard to what Corominas says, I take it he’s referring to relatively modern attestations in Arabic. Yule & Burnell make the caution: “The fact that the [post-medieval] Arabic dictionaries give a form ‘’sakirlat’’ must not be trusted to. It is a modern form, probably taken from the European word.”

According to Corominas:

‘Iskirlata se lee en una escritura árabe mudéjar de Zaragoza; es el nombre de unidad correspondiente al genérico ‘iskirlat (o ‘eskirlat), documentado en otras fuentes hispánicas o marroquíes, como Abensaíd (S. XIII), los Holal al-Mausiya (1381) y Pedro de Alcalá, y en el uso actual de Marruecos.

So there does appear to be some substantial support for the view that it is found in Medieval Arabic.  More generally, one cannot argue from the lack of surviving attestation that a word form did not exist (as my renowned former Secretary of Defense pointed out, “the absence of evidence is not the evidence of absence").  So I stand by my earlier statement that, even if the attestations cited by Corominas didn’t exist, “the fact that people in Arabic Spain were using iskirlata can be taken as strong evidence that it was being used in Spanish Arabic as well”.

Profile
 
 
Posted: 26 November 2010 04:36 AM   [ Ignore ]   [ # 14 ]
RankRank
Total Posts:  64
Joined  2010-11-02

If the attestation in Holal al-Mausiya (1381) is good and solid, it’s still nearly 300 years after the 1100 Latin scarlata. Pedro de Alcala (1505) is even later and I don’t forget that Pedro de Alcala’s Spanish-Arabic dictionary says that Spanish Almanac is called Almanac in Arabic, which I believe is a bogus translation because we know that the Arabs normally used words for an almanac other than the Almanac word. Abensaíd (S. XIII) is closer to 1100 but still more than a century too late, and we can’t see the exact wordform.

I tend to agree that the Concise OED is correct when it states that scarlet is “from medieval Latin scarlata, via Arabic and medieval Greek from late Latin sigillatus”, I just think that the word “probably” or “very probably” is called for. It makes a whole lot more sense than “Latin borrowed from Persian”.

Profile
 
 
Posted: 26 November 2010 04:44 PM   [ Ignore ]   [ # 15 ]
RankRank
Total Posts:  64
Joined  2010-11-02

Etymonline.com summarizes SCARLET as: “...from M.L. scarlatum “scarlet, cloth of scarlet” (cf. It. scarlatto, Sp. escarlate), probably via a Middle Eastern source (cf. Arabic siqillat “fine cloth"), from Medieval Greek and ultimately from L.L. sigillatus “clothes and cloth decorated with small symbols or figures...”

A thing I like about that statement is it allows for the possibility that the medieval Latin is not from a Middle Eastern source, while also conveying that the change in wordform from ‘’sigillatus’’ to ‘’scarlet’’ probably did happen within Arabic or another Middle Eastern source. The proposition ‘’sigillatus’’ ---> ‘’scarlet’’ is more plausible with an Arabic intermediary modifying the wordform.

Etymonline is also stating that ‘’sigillatus’’ ---> ‘’scarlet’’ is a certainty. That is too confident.

Profile
 
 
   
 
 
‹‹ Goat glanding      3rd person instead of 2nd ››