The Etymology of ‘Sabbath’
By Professor Francois de Blois, on 15 July 2015
The following is a lecture about the etymology of the ‘Sabbath’ given at the Workshop on ‘The Origins of the Seven-Day Week’ on 25 June 2015 at University College London. To download a pdf version please click here: SABBATH paper.
שַׁבָּת šabbòṯ m./f. ‘Sabbath, Saturday’ (Old Hebr. *šabbat)
שָׁבַת šòḇaṯ ‘to rest, cease’, (like Arabic sabata; also Ugaritic and Punic šbt in derived stems only)
שָׁבוּעַ šòḇūaʽ ‘week, group of seven’
שֶׁבַע šέḇaʽ ‘seven’
σάββατον n., σάββατα n. pl. ‘Sabbath, Saturday’, but also ‘week’
The more usual word for ‘week’ is ἑβδομάς f.
δὶς τοῦ σαββάτου ‘twice in the week’ (Luc. 18:12)
μία σαββάτων ‘Sunday’ (e.g. Matt. 28:1)
determined state: šabbṯā f. ‘Sabbath, Saturday, week’
absolute state: šabbā f., as in:
ḥaḏ b-šabbā ‘Sunday’
trēn b-šabbā ‘Monday’ etc.
(as opposed to trēn b-šabbṯā in Luc.18:12)
but Mandaic has šʼptʼ ‘Saturday’, presumably /šappṯā/
PERSIAN and PARTHIAN
The attested forms in Manichaean Parthian (Pa), Manichaean Middle Persian (MP), and in Early New Persian (NP):
Sunday: Pa. ēw-šambat; MP. yak-šambed, -t; NP. yak-šambih
Monday: Pa./MP. dō-šambat; NP. dō-šambih
Tuesday: NP. si-šambih
Wednesday: NP. č(ah)ār-šambih
Thursday: NP. panj-šambih
Friday: NP. āδīna; later normally Arabic jumʽa
Saturday: Pa. šambat; MP. šambed; NP. šambih; Judaeo-Persian also šambid
arḫu ‘moon, month, 1st day of the lunar month’
sebūtu ‘7th day of the month’
šapattu (or šabattu) ‘15th day of the month, full moon, fifteen days’
sebe f., sebet m. ‘seven’
sebū (also šebʼu, šabū) m., sebūtu f. ‘seventh’
The Hebrew name for the seventh day of the week, the Jewish day of rest, is, in the pronunciation implied by the Tiberian pointing, šabbòṯ, in Ashkenazic pronunciation /šabos/, in Sephardic pronunciation /šabat/; in Old Hebrew it was presumably *šabbat. It is used both as a masculine and as a feminine noun. Concerning the origin of this word there have been basically two positions: One is that it is a purely Hebrew word derived from the verbal stem š‑b‑t ‘to rest’; the other is that it is a loan word from Akkadian. We will look at these two etymologies in turn.
A connection between the verb šòḇaṯ ‘he rested’ and the noun šabbòṯ in the sense ‘day of rest’ is eminently plausible and the latter has explicitly been derived from the former at least since the time of Josephus (Ant. 1.1.1) who tells us that ‘in the dialect of the Hebrews’ the word σάββατα means ἀνάπαυσις. One does, however, need to consider the possibility that the noun is not derived from the verb, but rather that the verb is derived from the noun. It is true that in many of its occurrences in the Bible the verb š‑b‑t refers explicitly to rest on the Sabbath. Right at the beginning of the Bible, in the creation story, we read that on the seventh day God ‘rested’ (Gen. 2:2: way-yišboṯ) from his creative labour. But there are probably just as many passages where the same verb is used simply for ‘to rest’ or ‘to cease’ without any reference to the Sabbath, thus (a random example) Gen. 8:22: ‘While the earth remaineth, seedtime and harvest, and cold and heat, and summer and winter, and day and night shall not cease (lo yišboṯu)’. Both usages seem to be common in all compositional strands (J, E, P, D) of the Pentateuch, as well as in the other books of the Hebrew Bible. To maintain the notion that the verb šòḇaṯ is denominal from the noun šabbòṯ one would have to argue that the idea of the Sabbath is so central to Jewish thinking that a verb which originally meant merely ‘to keep the Sabbath’ became so engrained in the language that it developed into the general word for ‘stop’ or ‘cease’. But another difficulty is that this root has clear cognates in Semitic (Arabic sabata; also Ugaritic and Punic š‑b‑t in derived stems, all in the sense ‘to rest’). To explain this the ‘denominalist’ school has argued that Arabic sabata is either a loan from Hebrew, or that it has been influenced by Jewish usage, but this can hardly be the case with Ugaritic. On the other hand, the view that the noun šabbòṯ (*šabbat) derives from the verb š‑b‑t also has its difficulties, insofar as nouns of the patterns qattal, qittal, quttal are, as a rule, primary concrete nouns, often animal names (e.g. Syriac ṣepprā, abs. ṣeppar ‘small bird’), and not deverbal abstract nouns.
The concept of the Sabbath is inevitably tied up with that of the week, in Hebrew šòḇūaʽ, which is manifestly derived from the number ‘seven’ šέḇaʽ. The words for ‘week’ and for ‘Sabbath’ share the same first two consonants, but the third consonant is different: /t/ in the case of ‘Sabbath’ and ‘rest’, and /ʽ/ in the case of ‘week’ and ‘seven’. There is of course the view that Semitic or Hamito-Semitic roots originally consisted of only two consonants, to which a third, semantically empty, consonant could be appended (very much like the ‘root’ and ‘extension’ posited for ancient Indo-European), but this, if true, would apply to proto-Semitic or pre-Semitic but would have no relevance for a concept like that of the week, which emerged in historic times. We need therefore to regard š‑b‑t and š‑b‑ʽ as two totally separate roots.
The ancient Greeks did not have the concept of a week and thus there is no word for ‘week’ or for any of the days of the week in classical Greek. But in Greek writings by Jews and Christians we do have such terms. The seventh day is designated by the Hebrew or Aramaic loan word σάββατον (neuter singular) or σάββατα (neuter plural). Although there are a few passages where σάββατα does in fact mean ‘two or more Sabbaths’, in most cases both the singular and the plural forms are used to designate a single Sabbath. In particular, in the Septuagint we can observe that Hebrew šabbòṯ is translated either by σάββατον or by σάββατα without any discernable difference of meaning. It seems possible that σάββατα is in fact a borrowing of the Old Aramaic singular noun in the determined state *šabbatā (Middle Aramaic: šabbṯā), which Greek speakers subsequently reinterpreted as a neuter plural and that the sigular σάββατον is a back-formation.
The most commmon Greek word for ‘week’ is ἑβδομάς, a classical word for ‘group of seven’. But occasionally the ‘week’ is called σάββατον. This is clearly the case in Luke 18:12, the story of the pious tax collector and the hypocritical Pharesee, where the latter says that he fasts δὶς τοῦ σαββάτου, which must mean ‘twice a week’, not ‘twice on the Sabbath’. The names of the week days from Sunday to Thursday are formed from the cardinal or (more commonly) ordinal numbers plus the genitive σαββάτων, thus, for example, in Matt. 28:1 μία σαββάτων ‘Sunday’, which can be analysed as ‘(day) one of the week’, though I suppose it could also be understood to mean ‘(day) one of the (cycle of) Sabbaths’.
If we turn now to Aramaic, and specifically to Syriac, we find both the determined state šabbṯā and the absolute state šabbā. This word is manifestly borrowed from Hebrew, but with a decisive difference: In Hebrew the final /t/ is part of the root, but in Syriac it has been reinterpreted as the feminine ending /-t/, which in the absolute state is (here, as always in Aramaic) replaced by the ending /-ā/. Thus, while the Hebrew masculine or feminine noun šabbòṯ is attached to the root š‑b‑t, in Syriac the feminine noun šabbṯā belongs ostensibly to a root š‑b‑b, plus the feminine suffix -t.
The determined state šabbṯā is the Syriac word both for ‘Saturday’ and for ‘week’. The names for the days from ‘Sunday’ to ‘Thursday’ are formed from the cardinal numbers (the masculine forms for ‘one’ and ‘two’, but the feminine forms for the others) followed by the preposition b- and the absolute state šabbā. There is thus a distinction between trēn b-šabbā ‘(day) two in the week’, for ‘Monday’, and trēn b-šabbṯā ‘two (times) in the week’, which translates δὶς τοῦ σαββάτου in the above-mentioned passage in Luke 18. Friday is ʽrūḇtā ‘preparation (for the Sabbath)’, like Greek παρασκευή (notably Matt. 27:62). As can be seen, the Syriac designation of the days of the week is very similar to that in koine Greek, but I am not sure whether we are dealing with a Semitism in Greek or a Hellenism in Aramaic. I would note, however, that the constructions are not entirely identical: Greek follows the numbers with a genitive plural, but Syriac follows them by a preposition with a singular noun.
In Iranian languges we have names for the days of the week in Jewish, Christian, Manichaean and Islamic texts; the week is not used by Zoroastrians. The attested Parthian and Persian forms are listed in the handout. For ‘Saturday’ (and for the ‘Sabbath’ component of the compounded day names) we have šambed in Middle Persian and early Judaeo-Persian, and šambat in Parthian (and as a Parthian form in Middle Persian texts); these represent (I suggest) earlier and later borrowings of Aramaic šabbṯā (the former borrowed before, the latter after the Persian and Parthian voicing of post-vocalic voiceless stops) with dissimilation of ‑bb‑ to ‑mb‑, as in other Aramaic loanwords in Iranian (e.g. gumbaδ < qubbṯā, ‘dome’), and in the words for ‘Sabbath’ is many other languages (e.g. Geʽez sanbat, Church Slavonic сѫбота, OHG sambaztag > NHG samstag). The names for the days from Sunday to Thursday are formed (as in Aramaic) from the cardinal numbers followed by word for ‘Sabbath’; early New Persain āδīna is presumably the equivalent of ʽrūḇtā, but its etymology is contested. It is unlikely that the Persian Muslims had these names from the Manichaeans; it would seem rather that they borrowed them yet again from Christians or Jews. In any case, the Persian week is definitely an off-shoot of the Aramaic week. By contrast, the Sogdian Manichaeans used the planetary week.
I turn your attention now to the very interesting situation in Akkadian. The Babylonians and Assyrians did not have a week in the sense that we are using for this word, namely a fixed and repeating cycle of seven days, nor did they have anything really comparable with the Jewish Sabbath. They did however have several schemes for dividing their lunar months into periods of a fixed number of days, with their limits defined approximately by the cardinal phases on the moon: new moon, first quarter, full moon and third quarter. One of these schemes involves a set of special names for the first, seventh and fifteenth day of any month.
The first day is called arḫu (older warḫu), which is the ordinary word for ‘moon’, ‘month’, and then also ‘first day of the lunar month’, cognate with the word for ‘moon, month’ in most other Semitic languages.
The seventh day is called sebūtu, which is obviously derived from the word for the number ‘seven’: sebe (with a feminine referent) and sebet (with a masculine referent). These illustrate two significant linguistic events: First, the general rule that in Babylonian (the Southern dialect of Akkadian) all the Semitic laryngials except for /x/ are lost, in this case the final /ʽ/ of Semitic *s1abʽ. Second, the irregular replacement, in the word for ‘seven’, of Akkadian /š/ (for Semitic /s1/) by Akkadian /s/ (the usual reflex of Semitic /s3/), presumably as a result of dissimilation of the consecutive numerals ‘six’ (Semitic *s1idϑ) and ‘seven’ (*s1abʽ). sebūtu is formally identical with the feminine of the ordinal number sebū ‘seventh’; thus in principal sebūtu could simply mean ‘seventh’ with an implied feminine referent (perhaps mušītu ‘night, evening’), though the authors of the Akkadian dictionaries have preferred to see it as a separate, but homophonous feminine noun.
The fifteenth day is called šapattu or šabattu, which is also used for ‘full moon’ and ‘period of fifteen days’. As yet there is no satisfactory etymology for this. I suggest tentatively that it might be connected with the Akkadian word šaptu ‘lip’, cognate with Hebrew śòp̅ò, Arabic šafatun, Syriac sep̅tā, etc., from Semitic *s2ap-(a)t-. In Akkadian, besides meaning ‘lip’, šaptu has a number of figurative senses such as ‘edge, rim, bank of a river’; Syriac sep̅tā has similar usages. The full moon occurs when the sun and moon are in opposition; the moon rises when the sun sets and sets when the sun rises. It is thus conceivable that the Babylonians thought of the full moon as sitting on the ‘rim’ or ‘edge’ of the bright hemisphere. The proposed development of šaptu to šapattu would involve only a secondary gemination of the feminine suffix, something for which there are other examples in Akkadian. Although spellings which imply šapattu (with /p/) are more common, there are also spellings that unambiguously imply /b/, as well as many spellings that can be interpreted either way. Fluctuation between /p/ and /b/ is fairly common in Akkadian, and also in Babylonian Aramaic, including such tantalising examples as the Mandaic name for ‘Saturday’, šʼptʼ, presumably for šappṯā.
These three terms are mentioned together in at least two texts. They are not names of three random days, rather they belong together as designations for three important cultic events in any month. They also mark three of the cardinal points of the lunar month: the sighting of the new moon on the first day of the month, the first quarter on or about the seventh, and the full moon on or about the fifteenth. There is a longstanding discussion among both Assyriologists and Biblical scholars about a possible connection between the Hebrew šabbòṯ and the Babylonian šapattu. From the point of view of phonological correspondence the equation of the two words is not particularly problematic, especially if we take the Babylonian varriant šabattu (with voiced /b/) as our point of departure. šabbòṯ (older *šabbat) and šabattu have not only the same consonants, but even the same vowels; they differ really only in the distribution of the gemination. In late Babylonian the case endings were still written (often not correctly) but evidently no longer pronounced. Thus šabattu would have been pronounced as šabatt, but since Hebrew and Aramaic do not allow geminated consonants in final position the Hebrews would have reduced the final consonant to /-t/ and then perhaps compensated by geminating the labial in the preceeding syllable. The difficulty with this is the semantics. šabattu is the 15th day of the month, the time of the full moon, while šabbòṯ is the seventh day of a recurring cycle. Semantically it would seem actually more attractive to compare šabbòṯ with sebūtu, the seventh day of the month, but from a phonological point of view these two cannot very well be connected.
I offer the following as a hypothesis. The ancient Hebrews could have been aware of the Babylonian practice of celebrating a major festival on the seventh day of the lunar month, sebūtu. They might have adopted this as a name for the seventh day, then for the seventh after it, and the seventh after it, and so forth until the end of the month. Thus there would be four “sebūtus” in every month, on the 7th, 14th, 21st, and 28th day. The second sebūtu would coincide approximately with the Babylonian šapattu (šabattu) on the 15th day of the month, the day of the full moon. Prompted perhaps by the superficial similarity of sebūtu and šabattu, and by an association of both with Hebrew šέḇaʽ ‘seven’ (in the case of sebūtu correctly, in that of šabattu wrongly), and perhaps a further (spurious) association of both with Hebrew šòḇaṯ ‘to rest’, they might have come to regard sibūtu and šabattu as synonyms and applied the latter name to all four of the cardinal points of the month. We would then have to assume that at a later stage the Hebrews abandonned the practice of recommencing the counting of heptads at the beginning of every month and instead adopted a constantly recurring cycle of seven days without regard to the moon.
My proposal is thus to revive in a new guise the old suggestion that Hebrew šabbòṯ is a loan word from Babylonian šapattu. This means that it is etymologically not connected with the Hebrew root š‑b‑t ‘to rest’, thus relieving us of the mentioned difficulties of deriving either one of these from the other. The Sabbath is in essence the seventh day. The linguistic link with its seventh-ness is via the Babylonian sibūtu, which was contaminated (as we say in linguistics) with the unrelated šapattu. The connection with ‘rest’ is secondary, and essentially theological.
[This paper is for the most part a synthesis of the information contained in the standard dictionaries, in particular (for Hebrew) Koehler/Baumgartner3; Gesenius/Meyer/Donner (=Gesenius18); TWzAT; (for Greek) Liddell/Scott/Jones; Bauer/Reichmann/Aland/Aland (=Bauer6); TWzNT; (for Syriac) R. Payne-Smith; Brockelmann2; Brockelmann/Sokoloff; (for Mandaic) Drower/Macuch; (for Western Middle Iranian) Durkin-Meisterernst; (for Akkadian) CAD, AHw. I have not worked through the enormous mass of secondary literature, but I have benefited from re-reading Landsberger’s classic study Der kultische Kalender der Babylonier und Assyrer, 1915.]
 MacKenzie, BSOAS 31, 1968, p. 254. New Persian شنبه šambih )not šamba) rhymes with bih ‘better’ (see, for example, the verse by Niẓāmī cited s.v. in the Luγatnāma); the development of -eδ to –ih is irregular.
 Henning, Ein manichäisches Bet- und Beichtbuch, APAW 1936, 10, p. 85.
 See Encyclopaedia of Islam, 2nd edition, s.v. Taʼrīkh.