The Problem of the Etymology of "Rebetis" and "Rebetika"
by Ed Emery
The etymology of the words "rebetika" and "rebetis" has been a puzzle for half a century, and remains so.
In the course of work on a separate project (related to Arabic influences in and around Dante's Vita Nuova) I believe that I have found some possibilities of an answer.
Namely a derivation from the Arabic "ribaat".
This potential derivation seems not to have been examined in the available Greek literature on rebetika. This would seem to reflect the historically engendered and wilful ignorance of Greeks as regards Islamic, Ottoman and Arabic culture.
OTHER PEOPLE'S PROPOSED DERIVATIONS
Stathis Gauntlett, in his thesis (1978), went looking for a possible Turkish/Arabic derivation of "rebetis". He wrote: "The only instance of 'rebet' that I have found is in the 14th century Cypriot 'Assizes', where to judge by the context it appears to denote a type of imported liquor thus DuCange glosses the term: 'Vox Arabica, potus species'. I have not, however, found another attestation of 'rebet' as such in Arabic."
In my "Songs of the Greek Underworld" I list various other potential derivations that people have suggested over the years. [Note 1]
"Rembet", an old Turkish word meaning "of the gutter"; "rebenٍk" (pl. rebia'ta), a Serb word meaning "rebel"; "rab" (also reb, ruba'a, arba'a) Arabic and Persian words meaning variously "four", "quatrain", "God", "Lord"; "rab", a Hebrew word, from which derives "rabbi"; "remvastiko", a Greek word meaning "meditative", deriving from "remvo" or "remvazo", meaning "I wander". [Note 2]
There is also the well-known, sporadically occuring and habitually unsourced attribution to "rebet asker" ("an irregular soldier"), to which I shall return at the end of this paper.
PICKING UP THE SUFI CONNECTION
The rebetes habitually referred to themselves as "dervishes" and their hashish dens as the "tekke". Both of these are terms from the Sufi/Dervish tradition. It is worth examining whether the term "rebetis" might derive from that same cultural area. [Note 3]
This requires us to go back and look at history.
In the Dervish orders, the master was a revered sheikh. The pupil lived with him, shared his religious practices and was instructed by him. In times of war against the infidel, the pupil might accompany him to the threatened frontier and fight under his eye.
The term for this student who combined the religious life with the military was "murabit" ("one who pickets his horse on a hostile frontier"), derived from the root r-b-t. The Arabic "ribat" means "a frontier fort". By association with the militant religious orders, it then comes to mean a "monastery". [In post-Arabic Spanish this becomes "rapita", a term which still exists today in Spanish place-names.]
"Ribats were an Islamic creation small forts
in frontier areas occupied solely by warriors who had dedicated themselves for
a year or two to the defence of the faith." [The Different Aspects of
Islamic Culture, UNESCO,
The usage was already well established in the 11th Century, when Ibn
In Greek culture the qualities of the "rebetis" include poverty, basic human decency, a philosophical self-view, and a willingness to resort to armed action. These are the qualities that might have been found in the "murabit".
In more recent history the root r-b-t takes on further connotations. The Wehr-Cowan dictionary tells us that the Arabic word "ribaat" means inn for travellers, caravanserai, hospice (for Sufis or the poor).
I can provide no direct etymological link, bringing this word into the
Greek, but it seems to me that there is a very interesting case for deriving
"rebetis" from "ribaat" and "murabit".
Namely that the "ribaat" existed as a social
structure within the
A TANTALISING POSSIBILITY
There is a tantalising possibility here. For a while when I began this research I thought that the terms "ribaat", "murabit" were more characteristic of Western Islam (Spain, North Africa etc). I had not come across them with this Sufi/dervish meaning in Turkey/Asia Minor.
Therefore I hypothesised that it might have been brought across the
It is clear that the "ribat" is a historically and culturally specific institution. It is Arabo-Islamic in its origins. Therefore one would expect it to go into other languages and other cultures as a loan word, rather than being translated into local equivalents. [Note 6]
THEN THINGS BEGAN TO MOVE ON
Then I hit on a refinement which began to look promising.
As a testimony of the cultural durability of words deriving from the
root r-b-t we have a further development of usages that were apparently
widespread in the
Reckoning that a Turkish dictionary of the Ottoman period might produce
a result for "rebetis", I consulted "Redhouse's Turkish Dictionary"
(J.W. Redhouse, pub. Bernard Quaritch,
Redhouse gives the meanings with which we are already familiar, from the Arabic:
"Raabita": A band of union; system; regularity.
"Ribaat": A strong and secure place, formerly always fortified, where travellers, caravans, or military expeditions can take up their quarters on their journey, or on an enemy's frontier.
"Murabit": One who devotes himself entirely to the service of the faith, either as a warrior to guard the frontiers against external foes, or as a man of piety to pray for the welfare of the church and combat internal enemies."
However what is extremely interesting is the entry for the words "rabıtalı" and "rabıtasız" (given in Arabic script, which I cannot reproduce here).
The first, "rabıtalı", is given as meaning "Good. Capital. Excellent."
The second, "rabıtasız", is given as meaning "Bad. Not as it should be."
[The Turkish languages places "li" (or "lu" etc) after a word, to signify "with" something. So, if "sos" means "sauce", then "soslu" means "with sauce". Placing "siz" after a words signifies "without" something.]
Therefore "rabıta-lı" means "with rabita" or "having rabita". In formal terms (see above translation of "rabıta") this would mean "having union" or "having system" or "having regularity".
However I think, from the general context, that here we may be dealing with a widespread usage in Ottoman Turkish a phrase meaning "OK, fine, all in order, rock-steady" etc.
Redhouse does not set out to give rare usages in his dictionary (no
Sufi, no dervish, no tekke etc). His words are middle-of-the-road common
usages, chosen in his capacity as a fellow of the Ottoman Imperial Academy of
Sciences. If we find a word in Redhouse it is fair to assume that it has
currency throughout the
And by that token, it is reasonable to assume that "rabita" has a very everyday quality. It has "the quality of OK-ness... the quality of rock-steadiness... the quality of being regular".
The word "rabita" evidently has a fairly indefinable quality. If "rabıta-sız" (ie not-having-"rabita") means "bad, not as it should be", then this implies that "rabıta" also has a sense of "as it should be"... "comme il faut". And once again, if applied to persons it could reasonably mean "a proper person".
[My Langenscheidt Turkish dictionary obligingly provides "rabıtalı" as meaning orderly, well-conducted, level-headed person, coherent, consistent.]
Thus at this point, setting aside for a moment the possible Sufi, philosophical, dervish etc connotations examined above, it became tempting to think of "rebetis" as a Greek-derived word from an Ottoman Turkish usage in turn derived from the Arabic original of "ribaat". It could thus mean a person who has "rabıta". A "regular", "proper", "comme il faut" sort of person.
There is of course the problem of the implied vowel changes (a, i, e etc). How could "ribaat" transmute into "rebet"? In fact I think this presents no great problem. For instance the "ı" in "rabıta", the undotted Turkish "i", is pronounced as an indefinable "uh" and could very easily become an "e". The "a" (formerly a long "a") is another problem, which would have to be looked at. Certainly it is a relatively closed "a".
The adding of the "-is" to the end of the word is common Greek practice with Turkish words. For instance "tembel" (Turkish for "lazy") becomes "tembelis" (Greek slang for a lazy person).
[In passing, it would be tempting to think of "rebet" as one of those words which is massively present in popular culture, but almost entirely lacking from dictionaries and language-courses. Such as the Arabic kuwayyis, meaning nice, good, fine, pretty etc spoken on every street corner, but missing from the "official" language.]
I now turn to the additional associations of r-b-t with Sufi hostelry, dervish, inn, caravanserai etc, as noted above.
Call in at the bookseller's. He has dusty tomes of Arabic and Persian primers on a top shelf. At risk of my life I climb his little ladder and pull down a 100-year-old Persian dictionary. This attests that in 18-th century Persian "rebat" means an "inn, caravanserai, station for horses". [Note 7]
This is rather important. Several things have happened here.
First: We already know that the word Arabic-derived word "ribaat"
has at least a 500-year-long currency from the Western-most to the Eastern-most
shores of the Mediterranean... extending into Ottoman Asia Minor... With this
Persian usage attested, it now extends as far as
Second: As regards the word's pronunciation, we seem to have moved from the "ribaat" of the Arabic to the "rebat" of the Persian. A vowel change has taken place, which brings us tantalisingly close to the "rebet" of "rebetis".
Third: "Ribaat" is accented on the second syllable, between the "b" and the "t". "Rebetis" is also accented between the "b" and the "t". It feels as if this fact ought to be significant. I shall examine it further at another time.
Fourth: We might surmise that the "ribaat-rebat" was
one of the key social institutions of the
The ribat as "inn, caravanserai, staging post for horses" must surely have been one of the main places of public socialising available to local people. It was the place where travellers came and went and merchants displayed their goods...
We need historical accounts from the 18th-19th centuries, to be able to fill out the picture. I imagine that it was largely a male world. I wonder whether there was an association with prostitution. I wonder whether there was an association with hashish. I wonder whether people may have smoked and drunk. I wonder whether stories were told. I wonder whether people may have danced and sung. I do not have sources to confirm any of this. However, it might provide precisely the place and the ethos of the "rebetis" and "rebetika".
The best source for this would probably be the host of painters and
drawers who thronged to the
[Incidentally, we would also need to establish the relationship of the word Arabo-Persian word "ribaat" with the Turkish word "han", which also means "inn and caravanserai". Langenscheidt does not give "ribaat" as a Turkish word for inn, giving instead "han". But Redhouse is perfectly clear that "ribaat" meant inn, caravanserai etc under the Ottomans.]
Surely the word "ribaat" as "inn, caravanserai, staging post for horses" must have been entirely familiar to Asia Minor Greeks. And this opens a simple possibility:
"Rebetis" could be, purely and simply, a person who hangs out in a "ribaat".
[I should point out that there is an irritating and provocative research wrinkle here, as regards a possible derivation of "rebetis" from the Arabic "kharaba [Note 8].]
A DESCRIPTION OF A RIBAT FROM THE TENTH CENTURY AD
By way of
a diversion. My particular area of research is "Arabic influences in and
around Dante". So there I am, in the library, quietly minding my own
business, browsing through a book on "Gli Arabi in Italia" (Gabrieli
and Scerrato). In a close-printed and densely-worded Appendix there is a text
dating from AD 973 a description of the
Suddenly my eye lights on the word "ribat". And there, in a description from a thousand-and-something years ago, we have a definition of the reality behind the word.
"Giaccion su la spiaggia
A rough translation: "Along the shore of the sea, there are many ribats full of blackguards, low-life characters, people of a seditious nature, young men and old alike, full of fancy talk. They've got callouses on their foreheads from prostrating themselves, so that they can put themselves there and demand charity, and talk badly about honest women. Most of them are pimps for vile habits, or given to vices. They hang out in the ribats, like the people of no account that they are... people with no roof over their heads... a real rabble..." [Note 9]
So here, wonder of wonders, we have a detailed account of the kind of low-life characters who might hang out in a "ribat".
Leaving aside the calloused foreheads [Note 10], it comes pretty close to the ambivalent characterisation of the "rebetis". And that is in AD 973!
So, now we have the word "ribat" present as an
"inn-cum-staging-post" in the central
Having arrived at this point, I say once again that I have no direct etymological evidence for the transition of "ribat" into the Greek language as a potential origin of the term "rebetis". However the circumstantial evidence seems substantial. And certainly deserving of further analysis.
At this point I would return to the question of "rebet asker".
Commenting on an earlier version of this paper Kostas Vlisidis pointed me to an article [Note 13] in which the author, Kostas Karapotosoglou, argues (as I do) the derivation of "rebetis" from "ribat". This is encouraging.
However Karapotosoglou makes an odd assertion when, at the start of his article, he specifically negates any possible derivation from "rebet asker". To my knowledge, "rebet asker" is simply a Turkified equivalent of the Arabic "ribat askarii", meaning "a soldier of the ribat". In other words, the "rebet" which Karapotosoglou rejects is the same word as the "ribat" which he accepts. Would that we all knew Turkish and Arabic better than we do !
This paper is not intended to be conclusive. It is intended to open up terrain. If anyone has opinions on its contents, I would be very happy to hear from them.
Note 1: Songs of the Greek Underworld, trans. / ed. Ed Emery, Saqi Books, London 2000.
I should add that there is also great disagreement over WHEN the term "rebetis-rebetiko" first enters the Greek language.
In my book I cited the late Ole Smith (a stickler
for precision in research) to the effect that "it can be shown beyond
doubt" that the generic term rebetiko "made its first public
appearance as a musical term among the Greeks in the
Note 2: Sources, variously, Stathis Gauntlett, Costas Ferris and Gail Holst.
An obvious difficulty in deriving "rebetis" from "ribaat" would be if we accept the spelling "REMBetis" rather than "REBetis". However it is clear that when Greeks take a foreign word which has a "b" sound, it is rendered as "mb" (everything from "beer" to "bouzouki"). If the derivation is from "ribaat", it would certainly have come into written Greek as "remb ", "ρεμπ ".
Note 3: Stathis Gauntlett [personal communication] suggests that the
"dervish" and "Sufi" aspect cultivated in Rebetika may be
ironical. That may be so. However the title of the very well-known song
"Oli rebetes tou dounia" suggests, rather, an immersion in an area
that has Sufism etc as one of its coordinates. "Dunya" is,
apparently, a Koranic word, meaning not merely "people, mankind,
world" [Oxford Dictionary of Modern Greek, 1965], but the base,
mortal, earthly state of humanity prior to enlightenment. [Ibn 'Arabi, Felicitΰ
Note 4: Personal communication. And if one happened to have a taste for adjunct meanings, the similarly-rooted phrase "ribaatat al-ja'sh" means a person who is composed self-controlled, calm and intrepid. And "raabita" means link, bond, union (as in "raabita al-sadaaqa", bond of friendship).
Note 5: In response to a query, I received an e-mail from a Sephardi
Note 6: In the 12th century, the Italians found that they had no suitable word for the institution of the caravanserai-style inn, so they adopted the Arabic "funduq" (derived initially from the Greek "pandocheion") and turned it into the "fondaco" that one finds in Venice and Tuscany. They similarly derived their "arsenale" (a shipbuilding institution which they did not previously have) from the Arabic dar sina'a. Not to mention "tarifa", "dogana" ("tariff", "customs duty") etc. It is possible that the same mechanism applies, in bringing "ribaat" into Modern Greek being a foreign institution, for which no native word existed in the Greek.
Incidentally: In translating the "Ars Rhetorica" of Aristotle into Arabic, the Arab translators used "ribaat" to translate "syndesmos", meaning conjunction. "Syndesmos" also means the bonds of union that keep, for instance, a city together. It also means sodomy. These derivations appear to be of no particular use to us.
Note 7: An Iranian friend confirms that the usage "robat" as old-fashioned Persian, meaning
caravanserai, staging post particularly during the Safavid era (1502-1736).
Note 8: In Ottoman Turkish (cf Langenscheidt) "harabat" means (i) ruins; (ii) wineshops, taverns. Gauntlett (in "Rebetiko Tragoudi as a Generic Term") cites the Redhouse dictionary (1921 edition hence Ottoman usages) for "harabati" meaning "dissolute vagabond, especially a confirmed drunkard". It would be very tempting to tie all this up with "ribaat".
However (pending further dictionary work) it appears that all these meanings of "harabat" derive from the Arabic kh-r-b root ("kharaba" to destroy, wreck), and therefore NOT from the "r-b-t" of "ribaat". There would be no reason for Greek usage to drop the initial "h-a", since it derives from the "kh-a", a letter which would be available in Greek, as "chi-alpha" "χα" .]
A further interesting "harabat" connection was brought
to my attention by Hugo Strotbaum (personal communication). He cites Mark Slobin's Music in the culture of
"He [Sher Ali Khan] settled them and their families in a district of Kabul that was later termed the xarabat (from "xarab: ruined, debauched, indecent"; Steingass 1970: 451). The xarabat became the center both of lower-class musicians' dwellings and of Sufi (mystic) gathering places, a situation that prevails today..."
The "musician" and "Sufi" connections are tantalising here.
Note 9: Ibn Hawqal also makes reference to the "raabita
Note 10: Muslims when prostrating themselves in prayer touch the ground with their foreheads. Doing this to excess produces callouses on the forehead which the beggars show as a sign of their religious devotion.
Note 11: We need a way of assessing how common the "ribat" was, as an institution, as a staging-post for horses. A.W. Kinglake provides useful information in "Eothen" (1834):
"The distances between our relays of horses varied greatly: some were not more than fifteen or twenty miles; but twice, I think, we performed a whole day's journey of more than sixty miles with the same beasts." (p. 20)
This would suggest that on any given cross-country route there would be numbers of "ribats" strung out along the way possibly every 15-20 miles.
[Incidentally, this passage occurs in his description of crossing
In a useful passage Kinglake also describes the caravanserai:
"A caravanserai is not ill adapted to the purposes for which it is meant. It forms the four sides of a large quadrangular court: the ground floor is used for warehouses, the first floor for guests, and the open court for the temporary reception of camels, as well as for the loading and unloading of their burthens and the transaction of mercantile business generally. The apartments used for the guests are small cells opening into a kind of corridor which runs through the inner sides of the court." (In Gaza p. 162)
Note 12: Another Arab traveller, Ibn Jubayr [travelling in 1184-85], describes a ribat in the city of Ra's al-'Ayn: "The right-hand stream passes through a convent, also called al-ribat, built for Sufis and for foreigners, next to the spring." (Viaggio in Spagna, Sicilia..., p. 233)
In a later passage, etymologically interesting, he describes a ribat
It is important that he says "they are many in number". This was a widespread social institution.
Note 13: The article appeared in the Lexigrafikon Deltion Akadimies
Athinon, Vol. 16, 1986, under the title: "Synkritikes dierevniseis sta
Nea Ellinka: rebetis". It confirms that "ribat"
In the interest of widening the debate on my proposed derivation from "ribaat", I include below other illustrative materials relating to that term
1. "Ribat" has long had a sense of steadfastness and military
duty in the service of religion. The present-day Order of the Murabitoun
("Order of the Ribat") in the
"Be steadfast, hold each other to steadfastness, help each other to be firm (lit. make ribat) and have Towqah of Allah in order that you may be successful." [Qur'an 3:200]
2. In that light, the site at http://www.ummah.net/sos/caravan.htm contains the following reference:
"This is Ibn al-Mubarak, who used to perform the ribat for two months or more every year leaving aside his trade and the lessons of hadith, and going out for ribat, bewailing the fact that he has not performed ribat all his life, and that he has occupied himself with learning instead "
3. The present-day Palestinian National Authority is preparing studies
of the ribats to be found in
The description reads: "The sultan Qalawun al-Mansuri ordered the
construction of his pilgrim hospice for poor [Muslim] pilgrims to
4. For a major treatment of the ribat at
5. Similar photographic websites are also available for the ribat
6. The doctor Ibn al-Jazzaar (d. AD 980) lived and worked in Qayrawan
7. Kirsti Thorsen, lately engaged on post-graduate studies at King's College London, deals comprehensively with the possible relation between the rebetes and the world of Sufi in her dissertation "Mangas tha pei dervisis" (unpubl. 1999). I suspect that it is in this arena that the proof of my thesis will (or will not) be found.
8. Regarding Greek attitudes to Turkey, the following charming entry in D.N. Stavropoulos's Oxford Greek-English Learner's Dictionary (Oxford University Press 1989) is perhaps actionable under the Race Relations Act (as happened in the case of the the Babiniotis dictionary, which had to be withdrawn after its publication in Greece because of a racist reference to Bulgars):
bribe[ry], kickback, pay-off, graft: το λάδωμα δίνει και παίρνει στην Τουρκία, bribery /
graft is rife in
8. STOP PRESS: The Encyclopedia of Islam has a magnificent section devoted to the phenomenon of the "ribaat". This includes a note to the effect that, at a certain point and in some places, "ribaat" is co-terminous with "tekke". We are all aware of the rebetiko songs which cite the "tekke" as the focal point of rebetiko culture. Was "ribaat" co-terminous with "tekke" in the Greek communities of Ottoman Asia Minor? If so, that would bring me closer to proving my case...
9. DITTO: It is suggested that the elusive term "rebet asker"
has a derogatory aspect. It may be worth noting the following comment in
relation to the 1952 revolution in
10. The Treaty of
11. Kree Arvanitas points me to the social institutions of the
Jana Gough points me to the highly significant point raised in the Encyclopedia of Islam entry on "ribat": "It can be stated with confidence that to define it as a 'Muslim military monastery' is evidence of extrapolation and misinterpretation. It cannot be denied that the urban residences [my emphasis] of Sufis were subsequently known as ribat."
As regards the possibility of transition from "i" to
"e" (ribat to rebet), Leonidas Drisis points out that
another very similar (and well known case). It is the word 'sekleti' (worry,
frustration), which is also known as 'sikleti', and it appears with both forms
in several songs (which I cannot recall at the moment)".
[Paper presented at the Hydra Rebetiko Conference October 2001]