In this paper, the author compares the monophthongization of ai/ay and au/aw in Old Germanic dialects and Arabic dialects. The main question is whether the monophthongizations of ai and au into ē and ō in the Germanic and Arabic dialects are phonologically comparable. This turns out to be the case to a large extent. In both linguistic groups there were biphonematic diphthongs in the protophase, followed by a monophonematic period, after which monophthongization could occur. Moreover, in both groups there were dialects without monopthongization and other dialects in which monophthongization was not finished and intermediate forms like ei and ou occurred. Another question is whether contacts between Germans and Arabs played a part in mutual monophthongization. This turns out to be highly unlikely.
Intriguingly, a different monophthongization should be taken into account as well: au and ai into ā and a, which also occurs in both dialect groups. I have conformed to the Arabist terminology for distinguishing between these monophthongizations: monopthongizations into ē and ō are called contraction monophtongizations, those into ā and a they call elision monophthongizations.
Between the linguistic phases of Proto-Arabic and the modern Arabic dialects, many phonological changes have taken place. One of these changes is the monophthongization of ay and aw in various dialects. Ay monophthongized to ē or ī and aw to ō or ū. Both ay and aw could also monophthongize to ā (Iványi 2012).
Interestingly, a similar monophthongization is found in Germanic languages like Dutch, where the changes ai > ē and aw > ō have taken place. In addition, there are Germanic languages in which ai and au have been monophthongized to ā. Given these similarities between Arabic and older Germanic,1 I decided to investigate if the underlying phonematic processes of these monophthongizations in different language families were identical or not. I also wanted to examine whether they could have affected each other, which could be possible since a number of Germanic tribes, such as the Visigoths in Andalusia and the Vikings in Bagdad, had been in contact with Arabic-speaking people over various periods.
1.1 Monophthongs and Diphthongs
Monophthongs can be divided into short, half-long and long vowels. In Dutch, there are mainly short and half-long ones. Real long vowels occur only before -r, compare, e.g., pit—piet—pier [pɩt—pit—pi:r]. Classical Arabic has i, u and a as vowels with short and long versions.
Diphthongs generally consist of a full vowel plus a semivowel. The semivowels are i/j in the Germanic languages, y in the Arabic languages and w in the Germanic and Arabic languages. They often form a glide, a transitional sound (Iványi 2012). In Dutch, a glide occurs, e.g., in rooje from rode ‘red’.
When the diphthong ends in a semivowel, it is called a falling or closing diphthong; when it starts with one, it is called a rising or opening diphthong. Rising diphthongs occur, e.g., in Frisian, as in sjonge ‘to sing’, and Swedish, as in björk ‘birch’.
Languages can also differentiate between between short and long diphthongs. In short diphthongs, the full vowel is short; in long ones, it is long or half-long. In present-day Dutch, diphthongs are chiefly short: ei/ij as in leiden/lijden, ‘lead/suffer’ au/ou as in kauw ‘chew’ and kou ‘cold’, and ui as in lui ‘lazy’, and also ai and oi, only occurring in exclamations, like hai and hoi. Long diphthongs are aai as in lawaai noise’, ooi as in mooi ‘lovely’, oei as in loeien ‘moo, low’, eeuw as in leeuw ‘lion’, ieuw as in nieuw ‘new’, and uw as in huwelijk ‘marriage’ (Cohen et al. 1972). Modern Standard Arabic knows the short diphthongs ay and aw, as in bayt ‘house’ and law ‘if’, but long diphthongs are common enough in the dialects.
1.2 Biphonematic Diphthongs
Diphthongs are usually seen as one phoneme, so monophonemic or monophonematic. This applies to diphthongs in Modern Standard Arabic and in Dutch. Yet long diphthongs in Dutch ending in i/j are often interpreted as biphonematic, i.e. consisting of two phonemes (Cohen et al. 1972).
The Arabic diphthongs in this study were in any case biphonematic in Proto- and Classical Arabic, which means that the semivowels functioned as separate phonemes. The question remains whether they should be interpreted as consonants or vowels. As consonants, they would play a role in the radical structure; as vowels, they would not. The difference between consonants and vowels is also of importance within the syllable structure.
In a study of monophthongization, it is also important whether a biphonomatic diphthong can monophthongize at random. Does the process require an intermediate transition into monophonematic diphthong? And would the semivowel in such a monophonematic diphthong function as a consonant or a vowel?
1.3 Monophthongization in Germanic
Monophthongization as described above is also found in Germanic, because the diphthongs in Proto-Germanic were also biphonematic: ăĭ and ăŭ (Van Bree 1987, 100, 105). A question that may be raised here is whether there was a monophonematic intermediary phase and in what period the monophthongization took place.
1.4 The Study
In part 2, ‘Monophthongization of ai and au in Dutch’, I examine what has been written about this monophthongization in the Germanic languages, in historical grammars and historical-phonological descriptions of the Dutch language. I will also examine the changes in other Germanic languages.
In part 3, ‘Monophthongization of ay and aw in Arabic’, I explore publications on Proto- and Classical Arabic and modern Arabic dialects concerning this subject. Next, I turn to monophthongization in Egyptian Arabic. In doing so, I examine a number of specific dialects of Egyptian Arabic: Cairene and the dialect of Fayyūm. Cairene is the standard of Egyptian Arabic and reflects well how monophthongization has operated in an important modern Arabic dialect. In the dialect of the Egyptian oasis, Fayyūm, this monophthongization has not taken place.
2 Monophthongization of ai and au in Dutch
2.1 Monophthongization in the Germanic Languages
The Germanic languages feature a similar monophthongization as found in some varieties of Arabic. In Dutch, for instance, monophthongization took place of ai > ē and of au > ō. These diphthongs were biphonematic in Proto-Germanic: ăĭ and ăŭ (Van Bree 1987, 100, 105).
In order to gain an insight into this monophthongization, it is important to first examine the origin of ăĭ and ăŭ and their evolvement in the different Germanic languages.
2.1.1 Origin and Development of Proto-Germanic ăĭ and ăŭ in Germanic
Proto-Germanic (PGM) ăĭ and ăŭ evolved from Proto-Indo-European (PIE) diphthongs. The changes imply that these diphthongs were biphonematic, for the first part of the diphthong changed independent from the second part. In early PIE, the phonemes a and ā did not exist; in late PIE, they do occur (Beekes 1990, 160, 162). From here on, PIE means only late PIE.
PIE a > PGM a and PIE o > PGM a are general sound changes. In diphthongs the same happens to the first element: PIE ăĭ > PGM ăĭ and PIE ŏĭ > PGM ăĭ; PIE ăŭ > PGM ăŭ and PIE ŏŭ > PGM ăŭ. So the two PGM diphthongs ăĭ and ăŭ originate from four PIE diphthongs (Krahe and Meid 1969, 51, 53–54).
In the Old Germanic dialects/languages, these diphthongs evolved in various ways. Often monophthongization took place, but not always.
2.1.2 Development in the Various Old Germanic Languages
220.127.116.11 East Germanic
The Germanic language of which the oldest records exist is Gothic, now a dead language. The Goths originated in Scandinavia. In the first centuries
It is generally assumed that the phonetic value of ai and au in Visigothic was diphthongal if these graphemes go back to PGM ăĭ and ăŭ (which they not always do), but this is not certain. They may have had the phonetic values ē and ō, in which case they had been monophthongized (Van Bree 1987, 25–26). If they were diphthongal, they were in any case monophonematic diphthongs (Van Bree 1987, 100, 105). It is seen as implausible that they were monophthongized, considering the alternations which may occur within a paradigm, e.g., in taujan—tawida ‘make—made’ (Van Loon 1986, 46), but this argument is not totally convincing. There is a lot of evidence from loanwords in Provençal and Spanish that attest to the late survival of diphthongs in Visigothic (Dietz 1999).
About the extinct languages of the other East Germanic tribes such as the Vandals and the Burgundians too little is known for a statement about phonological changes (Van Bree 1987, 15).
18.104.22.168 North Germanic
Old Norse retained the diphthongs ăĭ and ăŭ as ei and au (Krahe and Meid 1969, 53–54). In other North Germanic dialects/languages, monophthongization did take place: ai/ei > ē and au > ȫ. Runic inscriptions indicate that this process started around 900 in Denmark and then moved up north. In central Sweden, mainly ai and au are found in the 11th century, although traditional spelling must be taken into account. But from the 12th century on, they are found no longer (Wessén 1968, 32).
According to Skard (1976), monophthongization came from Old Frisian and Old Saxon to the North. However, Old Frisian appears to have undergone a different process of monophthongization (see 22.214.171.124); it is unlikely, therefore, that this language played a role. Influence of Old Saxon, though, is possible: around 800, monophthongization was a fact in the northwest of Germany. In Danish and later in Swedish, the diphthongs ai and au ended up being monophthongized nearly everywhere, but in Norwegian only in minor parts. Icelandic and Faroese kept the diphthongs. As such, it can be stated that monophthongization is an East Norse phenomenon (Skard 1976, 57–58).
126.96.36.199 West Germanic
The West Germanic (WGM) languages show a diverse picture.
In Old Frisian, the monophthongizations ai > ā as in twā ‘two’, and ai > ē as in bēn ‘bone, leg’ took place. In addition, there was monophthongization of au > ā as in āge ‘eye’; when originally i or j followed in the next syllable, this ā changed into ē as in dēpa ‘baptize’ (Sjölin 1969, 20). In the changes ai > ā and au > ā, the palatalizing and rounding elements were elided and secondary compensatory lengthening took place. There has been a lot of discussion about the cause of the two different monophthongizations of ai, as Nielsen (1981, 130–134) points out.
In Old English, ai was also monophthongized into ā, as in gāt ‘goat’; au by contrast remained a diphthong, and evolved into ēa, as in ēare ‘ear’ (Krahe and Meid 1969, 53–54). Nielsen (1981, 134, 207) asserts that the evolvements ai > ā in Old Frisian and Old English took place independently.
According to Krahe, Old High German did not join in the monophthongization at all: ai > ei and au > ou. He also argues that, in Old Saxon—the dialect most closely related to Old Dutch—ai was always monophthongized to ē and au to ō (Krahe and Meid 1969, 53–54). As will be shown in 2.2.1 and 2.2.2, Krahe’s impression seems to oversimplify matters slightly.
2.2 Monophthongization in Dutch
Van Loon (1986) states that in all Germanic dialects there is a tendency to monophthongize au and ai, although the tendency is less or more distinct according to the district. He also points out that these au and ai are definitely no longer combinations of independent phonemes but are “real diphthongs” (Van Loon 1986, 46–47). He does not rate biphonematic diphthongs as real diphthongs. From his point of view, monophonematization precedes monophthongization.
The dialects of Dutch, Low German and High German can be regarded as a dialect continuum. Consequently, the developments concerning the monophthongization must be examined by combining evidence from all three dialects.
2.2.1 Development of au
In Old High German au was monophthongized before dentals and h (which Van Loon defines as a uvular blowing sound), as well as in word-final position, as in ōra ‘ear’, rōd ‘red’, hōh ‘high’ and strō ‘straw’. In other positions the diphthong remained as ou, as in ouga ‘eye’ and boum ‘tree’.
In Old Saxon and Old Dutch, au was only preserved before w, as in skauwon ‘witness, view’. In all other positions, au has been monophthongized. This process started around the year 700, as we can deduce from charters containing toponyms. In the East Limburg dialects, au shows the same evolvement as in High German (Van Loon 1986, 47–48).
2.2.2 Development of ai
Ai went through a more complicated development. In Old High German, it was monophthongized into ē, before r, w and h and at the end of the word, as in wē ‘woe’, and kēren ‘turn’. In other positions, ai continued to exist in a slightly contracted form as ei.
In Old Saxon, ai always became ē, except before a j-glide, as in WGM *aija > OS ei ‘egg’ and WGM *klaija > OS klei ‘clay’. In Dutch, on the other hand, ai also developed into ei before original i or j in the next syllable, as in Middle Dutch (MD) geite ‘goat’, eike ‘oak’ and beide ‘both’. For the rest, the evolvement was exactly like in Old Saxon, so steen ‘stone’, een ‘one’, weet ‘know, wit’ (Van Loon 1986, 49).
These observations hold for the central dialects of Middle Dutch. In West and East Flemish, as in Old Saxon, ai was more strongly monophthongized than in Standard Dutch (SD), e.g., MD clene ‘small’, eghen ‘own’, bede ‘both’, helech ‘holy’. East Limburg joins the Old High German evolvement, e.g., deil (SD deel) ‘part’, breit (SD breed) ‘broad’, ein (SD een) ‘one’, stein (SD steen) ‘stone’. In dialects in the provinces North and South Holland, and from there also in South African, occur vleis (SD vlees) ‘meat’, teiken (SD teken) ‘token, sign’ and deil ‘part’ (Van Loon 1986, 50–51).
Since the conditions were so complicated and because of the influence of various dialects and analogy on Standard Dutch, the present alternation ē—ei does not match the rules provided above. SD features verbreden ‘broaden’ as well as verbreiden ‘spread’, waarheden ‘truths’ as well as waarheid ‘truth’ (Van Loon 1986, 50–51), gereed ‘ready, done’ as well as bereid ‘ready, willing’ and eekhoorn ‘squirrel, literally oak-horn’ as well as eik ‘oak’ (Van Bree 1987, 101), heem ‘farmyard, home’ as well as heimelijk ‘secret(ly)’, steen ‘stone’ as well as -stein in names of castles (Van Loey 1970, 76).
3 Monophthongization of ay and aw in Arabic
3.1 Developments in Arabic in General
3.1.1 Diphthongs in Proto-Arabic and Classical Arabic
The Proto-Semitic diphthongs ay and aw have been preserved in Proto-Arabic. They may therefore have existed in Proto-Arabic of old, but they may also have arisen in other ways, such as the shortening of the long diphthongs āy and āw (Iványi 2012).
Since y and w are radical consonants,2 the question arises whether genuine diphthongs are possible in Arabic, but there are examples which prove that the possibility exists. Contracted forms like ṯawbbakrin and jaybbakrin from Classical Arabic and a modern Lebanese form like mawtna ‘our death’ prove that the -w- and the -y- cannot be interpreted as consonants here.3 The y and the w are therefore vocalic elements and the diphthongs take the position of a long vowel in the syllabic structure (Fleisch 1961, 147). Fischer (1967, 61, 73) draws the same conclusion about this subject.
3.1.2 Diphthongs and Monophthongization in Modern Arabic Dialects
In a large number of the Arabic dialects the diphthongs have been monophthongized, but often alongside the new monophthongs some diphthongs have also been preserved. Sometimes monophthongization did not occur, e.g., in some of the Lebanese and Yemeni dialects and Maltese (Diem 1985; Iványi 2012).
In the eastern dialects, there has usually been monophthongization of aw > ō and ay > ē and in most western dialects of aw > ō > ū and ay > ē > ī. There ō and ē have been raised secondarily. So in the eastern dialects five long vocals have evolved, in the western dialects initially also five, but there no more than three survived (Iványi 2012).
On the Arabic peninsula dialects exist in which diphthongs are completely absent, while in others they do appear. In the region of Nuṣayri (in Syria) the alternation aw/ay—ā is found, so the monophthongizations aw > ā and ay > ā. In Iraq the monophthong ē that evolved from ay can be broken, as in zēn > z(i)yēn (Iványi 2012). The result of this breaking is a rising diphthong.
In the western Arabic dialects three different types of diphthongs can be distinguished, the third of which is a secondary diphthongization by breaking, i.e. a long vowel breaking into two elements. The products of monophthongization in these dialects are usually ū and ī, but ō and ē also occur. Not uncommonly the diphthongs have been partly retained: aw > ōw and ay > ēy or aw > ou and ay > ei. Such partly retained diphthongs are found with nomadic peoples, but also in the northern parts of Algeria and Morocco, and Malta (Iványi 2012).
3.1.3 Period of Monophthongization
It is not established when the monophthongization took place in Arabic. Diem (1985) has examined whether this could have taken place as early as the 1st century
Another possibility would be to examine loanwords from Arabic into other languages. However, from the spelling of those words it is not possible to deduce the pronunciation with certainty. As such, no conclusions are possible: there is a possibility that monophthongization had already taken place in the dialects but that this was not reflected in the spelling. In this early period we are talking particularly about loanwords into Greek, that is: the Greek spelling of Arabic names from papyri and inscriptions, e.g., Μελεχ = Milēk < Mulayk, but it is also possible that Μελεχ = Melek < Malak. Sometimes there are strong indications that monophthongization is under way, but there is no certainty, since it concerns only individual cases, in which it is not always clear what name is referred to. The presumption is, for instance, that Λος is the indication of l-ōs < al-Aws (Diem 1985, 76–77). Diem does not indicate where these Greek names were written. There are hundreds of examples, in the papyri. It is fairly obvious that the variety that spread with the Arab conquerors had preserved full diphthongs. Greek was just not very good at rendering them.
The borrowings into Arabic are even more problematic. A large part of early loanwords reached Arabic via Aramaic, which had its own phonetic laws. For instance, Aramaic had grammatical change, a change within the paradigm: aw—ō and ay—ē, e.g., baytā—bēṯ and yawmā—yōm. This makes conclusions about the phonetic value in Arabic extremely difficult (Diem 1985, 77).
There is, however, one case to which these difficulties do not apply. The Byzantine stronghold in Fusṭāṭ, Egypt, where from about 643 the centre of the Islamic authority had its residence (Jomier 2012), bore the name of Βαβυλων. In Arabic papyri, this name is twice spelled Bāb-il-yōm ‘door of the day’, an abundantly obvious case of folk etymology. From the Arabic spelling, it is indeed not possible to gather whether it reads Bāb-il-yawm or Bāb-il-yōm, but the Greek word Bāb-il-ōn with long ō could never have resulted in the folk etymology Bāb-il-yawm with aw. Consequently, the Egyptians must already have had the monophthong ō (Diem 1985, 77–78). However, that conclusion does not necessarily follow; it is possible that the Egyptians heard Bāb-il-ōn as Bāb-il-awn and interpreted this as Bāb-il-yawm!
Werner Diem is of the opinion that with this, combined with the Greek spelling of Arabic names, he has found the evidence that in the Arabic dialects of Egypt the monophthongization of aw > ō (and with that also of ay > ē) took place in the 1st century of the Hijra (Diem 1985, 78).5 This conclusion, however, is completely unconvincing.
3.2 Development in Egyptian Arabic
The Arabic diphthongs ay and aw evolved in the eastern dialects mostly into ē and ō (see 3.1.2). That holds also for Egyptian Arabic, which falls for the greater part under the eastern dialects, but partly also under the western ones (Behnstedt and Woidich 1985).
Sometimes there is a secundary pausal diphthongization: ē > e.i and ō > o.u (Behnstedt and Woidich 1985, ii, map 2). This is not consistent, but occurs mainly in the Daqahliyya, a government in the eastern Delta, and then mainly with female speakers and moreover mostly in frequent words like ey ‘what?’ and ley ‘why?’ (Behnstedt and Woidich 1985, I, note on map 2).
Especially in a number of oases, like Farāfira, Dāxila, Banī Swayf and Fayyūm, but also in the western coastal strip along the Mediterranean and in part of the western and eastern Delta diphthongs still occur, which are sometimes monophonematic and sometimes biphonematic. In these regions, partly western Arabic dialects are spoken (Behnstedt and Woidich 1985, ii, map 1).
The Arabic diphthongs ay and aw evolved in Cairene into ē and ō. Nevertheless, ay and aw still exist in a few purely dialectical words and in a few Standard Arabic words commonly used in the ordinary language (Norlin 1985, 244), like law. Ay and aw have also been preserved in words like mawgūd and awlād, in which after monophthongization the word structure and the root would no longer be transparent, as the w is one of the three radicals in these words (De Jong 1992, 44). In this modern Egyptian dialect, the diphthongs ay and aw have often been preserved through gemination of the semivowel, which is then usually spelled double, e.g., ṭayyib and ʾawwil (Blanc 1981, 193–194).
According to Blanc (1981), the development of ay > ē and aw > ō was a gradual process and did not happen with the same force everywhere. According to him, at the end of the 19th century there were still a few districts of Cairo where people said ay and aw, but in fact the common Egyptian pronunciation ē and ō had been firmly established around 1830. We must bear in mind that Blanc bases himself on Judaeo-Arabic sources (Blanc 1981, 195). Judaeo-Arabic can vary from the standard dialect (Rosenbaum 2003). In the 18th-century Hebrew transcriptions he examined, aw or ō do not occur, but both ay and ē are present,6 from which can be concluded that a beginning monophthongization is in progress. From 17th-century sources not much can be deduced concerning this matter (Blanc 1981, 195–196).
In Cairene, there are three specific words with an u—aw variation in the declension. One of these is ḥuṣla/ḥuṣála/ḥuṣāla ‘chicken’s crop’, with ḥawāṣil as plural. From the plural can be deduced that the first syllabe of the singular must have contained the long vowel ō or the diphthong aw, so *ḥōṣala or *ḥawṣala. As the a in the second syllable has to be stressed and is therefore preserved, ḥuṣála is the form to be expected according to phonetic law, and in addition to this, other forms have arisen, for which Woidich gives an explanation (Woidich 1991, 1637).
When the pseudo-duals7 riglēn, ʿenēn en ʾidēn, in which ē has been monophthongized from ay, are followed by a possessive suffix, the -n is dropped. With the possessive suffix of the 1st person singular -ya, the -ē under the influence of -ya becomes -ay, so riglayya, ʿenayya and ʾidayya. So, another diphthong has arisen (Woidich 2006, 122). As a matter of fact, it is also quite possible that the -ay in these words has never evolved into -ē but always remained -ay.
Like other long vowels, the ē and ō evolved from ay and aw are, by force of certain conditions, shortened to i and u, as in (Woidich 2006, 31):
A long ē which loses its stress or is followed by two consonants, is shortened, and therefore usually replaced by i, as we have just seen. Yet there are speakers of Cairene who do not replace ē by i, but by a. This happens in the perfect of certain verbs when affixes are added, as in:
The alternation ē—a also occurs in itnēn ‘two’ as opposed to itnanāt ‘in twos, by twos’ and itnantēn ‘two twos, two pairs’. In Cairene this alternation is rather unexpected, but in the rural dialects it occurs frequently. The alternation ē—i is the common one in Egypt in general. The only exception is the dialect of the Baḥariyya oasis, where the alternation ē—a is the rule; there they have a similar rule for ō, the alternation ō—a (Woidich 1991, 1643–1644).
There is a group of words in Cairene in which a occurs where according to the rules of Cairene Historical Phonology one would expect i. For example zatūn < *zaytūn and lamūn < *laymūn. In these cases instead of a we would expect ē, which would then have evolved into i (Woidich 1991, 1645). For an explanation we have to take the older form ay. Dialects which have kept the diphthongs, often monophthongize ay by elision of the glide (the semivowel y), as is shown in a few variants of present-day Arabic. The result is ā, with compensatory lengthening for the loss of the glide, or a, when unstressed long vowels are shortened in the dialect. This happens, for example, in the Egyptian country dialects in the provinces Fayyūm and Banī Swayf, where the diphthongs were preserved. A few examples:
The main stress is on the last word of this compound. Woidich supposes that the same reduction of ay must have taken place in an earlier stage of Cairene, when the diphthongs still existed. He demonstrates this, e.g., on the basis of the example báyt—baytáyn. By elision of the glide the first ay of baytáyn was monophthongized into a and so batáyn arose. After that the ay of báyt and the ay of batáyn were monophthongized into ē, with the result bēt alongside batēn. So diachronologically this is easily accounted for, but synchronically it is not transparent, it does not agree with the common rules. This explains why in the end bēt—batēn was restructured into the set bēt—bitēn. Even so, some speakers of Cairene have kept the pronounciation a from the earlier stage, so batēn. So with these speakers the alternation ē—a is a surface alternation, it is not grounded on phonological rules.
Woidich concludes that the explanation of this alternation that is still found with a few speakers of Cairene, together with its parallel, the alternation ō—a, is crucial in the morphophonology of the Baḥariyya dialect mentioned before (Woidich 1991, 1645–1648).
3.2.2 The Fayyūm Dialect
As stated before, in Fayyūm and the nearby Banī Swayf diphthongs are still present, which are generally seen as sometimes monophonematic and sometimes biphonematic (see 3.2). Map 1 of the dialect atlas of Egyptian dialects shows that the old diphthongs ay and aw in that area are pronounced as follows: ay as [ei] and aw as [ou], but after the pharyngeals ḥ and ʿ as [ai̭] and [aṷ] (Behnstedt and Woidich 1985, ii, map 1). Woidich also mentions in an article about Egyptian Arabic in historical perspective, that the diphthongs ay and aw have been preserved in these areas, as opposed to in Cairene, which may be seen as Standard Egyptian. Speakers from these areas know that Cairene ō matches with their own aw. This means that all words with ō taken over from Standard Egyptian acquire aw, even loan words which historically have never had a diphthong. Examples: ʾawḍa for ʾōḍa ‘room’ (from Turkish oda) and ṣawbar for ṣōbar ‘fertilizer’. This aw is treated like historical aw and reduced to a before two consonants: ʾaḍt in-nawm ‘bedroom’ (Woidich 1997, 186–187), as we have seen in 3.2.1.
De Jong (1992, 6–7) states that the most striking phenomenon in the Fayyūm dialect is the presence of the diphthongs ay and aw, for which Cairene has respectively ē and ō, and that apart from that the monophthongizations (shortenings) of these diphthongs are an important characteristic. The allophones of ay have values from [ai] to [ei], but the influence of ḥ and ʿ and the influence of emphatics and q give rise to slight pronounciation variants. The same goes for the allophones of aw. They have values from [au] to [ou] and [o:], but the influence of the phonemes mentioned before also changes the pronunciation slightly.
As stated above, the ō in loanwords from Cairene is realised in Fayyūm as aw. Thus Cairene banṭalōn becomes banṭalawn/manṭalawn ‘trousers’ en mutōr is changed into matawr ‘engine, motor’. The ū which does not originate in loan words is also often diphthongized into aw or ow, as kawsa from kūsa ‘courgette’ and fūl sowdāni from fūl sūdānī (De Jong 1992, 58).
The word used in Fayyūm for carbon paper, however, is waraʾ karbūn and not waraʾ *karbawn. This is caused by the fact that the word is almost only used in schools and therefore is not influenced by a “lower” language level (De Jong 1992, 8–9).
Both ay and aw can be monophthongized through shortening: ay > [a] and [e]; aw > [a] and [o]. The shortenings happen when the diphthong is followed by two consonants or does not have the primary stress (as also in 3.2.1). De Jong (1992, 8–10, 41–42) adds that also rapid speaking and/or carelessness may be of influence.
As stated, in Cairene occur the forms zatūn < *zaytūn and lamūn < *laymūn (see 3.2.1). In Fayyūm zaytūn and laymūn occur, but in rapid speech also zatūn and lamūn. The Fayyūm word ʾaburž, which stands for the ‘Auberge (du Lac)’, may be explained in the same way. It must have been borrowed initially from French as ʾōbirž. Since the ō in loan words was interpreted as aw, the word must have changed into *ʾawbirž. Then the i was labialized to u, by which the word became *ʾawburž. Finally the shortening of aw in unstressed position set in and the result was ʾaburž (De Jong 1992, 55).
The shortening of ay and aw happens through elision of the second component of the diphthong. The monophthongization can be cross-word-border and does not only apply to morphologically formed diphthongs, but also to non-morphologically formed ones. An example of a shortened morphologically formed diphthong is xamas baḍāt from bayḍāt. A non-morphologically formed diphthong can produce the following shortening: šwat ḥagna (kida) from šuwayyit ḥagna via šuwayt ḥagna (De Jong 1992, 42–43).
Just as forms like mawgūd and awlād can in Cairene not be monophthongized into *mōgūd and *ōlād, these words can in the dialect of Fayyūm not be monophthongized into *magūd and *alād (De Jong 1992, 44).
On the basis of several indications De Jong (1992, 58) has concluded that the diphthongs ay and aw in Fayyūm must be seen as monophonematic. His arguments are very convincing. De Jong states that the Fayyūm dialect is in a transitional phase as far as diphthongs are concerned: synchronically and in certain positions they are being monophthongized. He also states: “The diphthongs, however, have already been monophonematized” (De Jong 1992, 2). This implies that from his point of view monophonematization preceded monophthongization.
4 The Research Questions
The research questions are as follows:
- Are the processes behind the monophthongizations here discussed, which occur in different language families, the same or not?
- In what era did the monophthongizations happen? If the eras match, could there be some connection?
In order to answer these questions, not only the monophthongization of ay/ai en aw/au into ē and ō should be considered, but also that of ay/ai and aw/au into ā. The first two are contraction monophthongizations, the other two elision monophthongizations.
4.1 Similarities and Differences between the Arabic and Germanic Monophthongizations
There are a number of similarities and differences between the Arabic and Germanic monophthongizations.
4.1.1 Similarities between the Two Dialect Groups
1) The diphthongs ay/ai and aw/au were originally biphonematic (see 1.2 and 2.1).
2) Monophonematization preceded monophthongization. The relevant literature does not comment on this explicitly, but it is strongly suggested, e.g., in the deductions of De Jong and Van Loon (see 2.2 and 3.2.2).
3) The second element of the diphthong could be interpreted as a vowel. For Arabic there are examples which prove that the possibility exists (see 3.1.1). In the literature about Germanic, the diphthongs within the syllabic structure are put level with long vowels, so implicitly it may be assumed that the second element is a vowel.
4) The contraction monophthongizations of ay/ai and aw/au into ē and ō occur for the greater part in both dialect groups (see 3.1.2, 2.1.1 and 2.1.2).
5) In a number of dialects of both groups a partial contraction has taken place, in the course of which a diphthong or a diphthongic monophthong was retained.
In Arabic this goes, e.g., for the partly preserved diphthongs in the dialects of some nomadic peoples: aw > ōw and ay > ēy or aw > ou and ay > ei (see 3.1.2). In Fayyūm ay is pronounced as [ei] and aw as [ou] (see 3.2.2).
As to Germanic: in Old High German, the transitions ai > ei and au > ou have nearly always taken place. In Old Norse ai > ei is found and in Dutch ai under certain conditions became ei (see 188.8.131.52, 2.2.1 and 2.2.2).
6) The elision monophthongizations ay/ai and aw/au into ā occur to some extent in both dialect groups. After the elision of the glide, a became ā by compensatory lengthening. These monophthongizations are found in a number of versions of Arabic in which contraction monophthongizations do not take place (see 3.1.2, 3.2.1 and 3.2.2), but also in Germanic dialects like Old English (only with ai) and Old Frisian (both with ai and au) (see 184.108.40.206).
4.1.2 Differences between the Two Dialect Groups
1) The outcomes of monophthongization ē, ō and ā have only partly remained in both dialect groups. In some dialects, they have undergone further evolvements. Those further evolvements are different in both groups.
Thus in the western dialects of Arabic ō and ē have been secundarily raised to ū and ī. And in Egyptian Arabic we find, under certain circumstances, shortening of ē, ō and ā (see 3.1.2, 3.2.1 and 3.2.2).
In North Germanic au > ō has changed into au > ȫ. Old English ā (< ai) later often has become oa, as in *gait > gāt > goat. Old Frisian ā (< ai) has in modern Frisian often become ea, as in *gaist > gāst > geast; Old Frisian ā (< au) has become [iə], written as ea, as in *baum > bām > beam.
2) In the literature about Arabic, nothing is to be found about any central area from which the monophthongizations of ay and aw to ē and ō could have spread. For Germanic, this appears to have been Old Saxon. It is in this dialect that the monophthongizations of ai > ē and au > ō first took place and where they are most salient. From there, the phenomenon must have spread. In Dutch, there is less monophthongization of ai and in High German there is even less monophthongization of both ai and au. From Old Saxon the monophthongizations moved northwards, first into Danish and then into Swedish and partly into Norse. The peripheral dialects in Sweden were never reached, nor the peripheral island languages Faeroerese and Icelandic.
3) An important difference between the perception of monophthongization in both dialect groups is that Arabic-speakers are still faced with the monophthongization: they find diphthongs in the Koran and in Modern Standard Arabic, while in their own dialect the same words often have monophthongs. This may suggest to them that those monophthongs may originate from the diphthongs.
In Germanic, the monophthongizations have taken place before the different dialects evolved into languages. The speakers of the present Germanic languages have no idea that the diphthongs in, e.g., High German or Icelandic words are more original than the phonemes in the corresponding words in their own language.
4.2 Is There a Connection between the Monophthongizations of
ay/ai and aw/au in Arabic and Germanic?
Can the diphthongs referred to in this study in both dialect groups have influenced each other? Such an influence is only possible if close language contact existed and if the monophthongizations took place in the same period.
4.2.1 The Dating of the Monophtongization
For Germanic, the dating of the change is as good as established. In 4th- and 5th-century (Visi)gothic there was probably no monophthongization (see 220.127.116.11). In the West Germanic languages, monophthongizations (into ē and ō) started around the year 700 and in the North Germanic languages a few centuries later (see 18.104.22.168 and 2.2.1).
Determining the period when monophtongization took place in Arabic is much more complicated, especially given the conservative orthography. Diem suspects that monophthongization in Egypt may already have taken place in the 1st century
4.2.2 Language Contact
The other criterion is the close language contact. The Arabs had contact with Visigoths and Vikings on the Iberian peninsula as well as with Vikings in Russia and the countries around the Caspian Sea.
The Vandals, but especially the Visigoths, played an important role on the Iberian peninsula from the beginning of the 5th century. After centuries of devastating battle—first with the Vandals and then internal—the Visigoths gained the power over Hispania (Menocal 2006, 27). Nevertheless, they continued to be a minority (Menocal 2006, 30). The role of the Vandals was played out by then.
In 711, Ṭāriq ibn Ziyād crossed the Straits (Gibraltar has been named after him: Jabal Ṭāriq), determined to enter into battle with king Roderic of the Visigoths. He defeated him at Wādī Lago (Rio Barbate). Within a year he had conquered the whole of Spain (Lévi-Provençal et al. 2012). The last kings of the Visigoths were corrupted and decadent and the North African muslims had no trouble at all capturing their kingdom (Menocal 2006, 28). So, the Arabic speakers made close contact with the Visigoths, but this contact was not of a character which invited them to copy linguistic developments like monophthongization from each other. This would not have been possible anyway, as it is not likely that monophthongization occurred in Visigothic, and most probably it had not yet occurred in North African Arabic either.
22.214.171.124 The Vikings
From the 9th century onwards, the Arabs repeatedly had to deal with the Vikings. In Spain they were Norwegians and Danes. Their first attack on Islamic Spain took place in 844, and in the following centuries many more confrontations followed. The Vikings in the West were called al-Majūs, those in the East al-Rūs (Melvinger 2012).
The eastern Vikings were mainly Swedes. They sailed southward by way of the Baltic Sea and Russian rivers as far as the Black Sea and the Caspian Sea and surroundings. Many of them settled in Russia for a long time and got to be held in high esteem. The local population called them Rus: Russia owes its name to them (Almgren et al. 1975, 133). The Rus traded with the Byzantine Empire, and with Arab merchants and trade missions along the Volga, e.g., in 921 with the diplomat ʾAḥmad ibn Faḍlān from Baghdad, who described Nordic merchants (Almgren et al. 1975, 134, 138–139). They also got into contact with Persians and Arabs near the Caspian Sea, e.g., in present-day Iran and Iraq. This Islamic area they called Serkland/Särkland. Runic inscriptions in Sweden testify that they have been there, and enormous treasures of Persian and Arabic silver coins in Scandinavia also furnish proof (Almgren et al. 1975, 139).
In that period—between about 850 and 1050—the North Germanic languages had hardly experienced any monophthongization; only among the Danes there were beginnings of monophthongization from about 900 (see 126.96.36.199). Nothing is known, actually, about monophthongization in the same period with the Arab peoples they traded with. The only thing is Diem’s hypothesis, which is restricted to Egyptian Arabic (see 3.1.3). As such, there is no evidence for mutual influence.
We must conclude that in the Modern Arabic dialects there is a strong tendency to monophthongize the Proto-Arabic diphthongs ay and aw. A similar tendency can be observed in the Old Germanic dialects about Proto Germanic ai and au. This similarity stirred me to study the questions I had a closer look at in section 4.
In both dialect groups the diphthongs were originally biphonematic. The most important monophthongizations are the ones of ay/ai and aw/au into ē and ō, the contraction monophthongizations. For these, Maṭar (1967, 69–70) and ʿAbd al-Tawwāb (2000, 50–51) use the term ʾimāla: the sounds assimilate, they shift towards each other. To Arab grammarians and Arabic linguists, this term usually means the raising of ā to ē or ī, but here it is used in a more literal sense. In addition, we find the monophthongizations of ay/ai and aw/au into ā, the elision monophthongization. Both types of monophthongization occur in both groups.
It is not entirely clear when the monophthongizations took place. Little has been written about the subject for the Arabic dialect group and, so, more research is necessary. For the Germanic dialect group, the contraction monophthongizations are dated from 700 onwards (see 4.2.1).
In both dialect groups, dialects exist in which the monophthongizations have not or only partly taken place. In Arabic, the diphthongs have been preserved, e.g., in Fayyūm and partly preserved in regions like northern Algeria: aw > ōw and ay > ēy or aw > ou and ay > ei (see 3.1.2, 3.2 and 3.2.2). In Germanic, the diphthongs are, e.g., still found in Icelandic; ou and ei in Old High German and a few other languages can be seen as not completely executed monophthongizations (see 188.8.131.52, 184.108.40.206, 220.127.116.11, 2.2.1 and 2.2.2).
The similarities between Arabic and Germanic monophthongizations—in both groups biphonematic diphthongs are the basis, in both groups there are similar contraction and elision monophthongizations, etc.—are much more substantial than the differences (see 4.1.1 and 4.1.2). Moreover, the differences are mainly secondary, e.g., the development of the phonemes after the monophthongization and the perception of monophthongization. As such, the underlying processes are likely the same.
For the rest, it must be concluded that the monophthongizations in the two dialect groups have come about independently, that they have not influenced each other, even if they may have taken place, perhaps partly, in the same era. The fact is that the Germanic peoples with whom the Arabic-speaking people had contact—the Visigoths in the 8th century in Spain and the Vikings between 850 and 1050 in Spain, Russia and the regions near the Caspian Sea—had experienced hardly any or no monophthongization. Therefore they cannot have had any influence. In all probability, the reverse is equally true: there has been no influence from the Arabic side. Whether they did already have monophthongization is not known, and also the contacts were not in such a way that the languages could have affected each other (see 18.104.22.168 and 22.214.171.124).
* I would like to thank Peter Alexander Kerkhof, Thijs Porck and the anonymous reviewer for valuable remarks and suggestions for the content of this paper.
Behnstedt and Woidich 1985: P. Behnstedt and M. Woidich, Die ägyptisch-arabischen Dialekte, Band i: Einleitung und Anmerkungen zu den Karten, Band II: Dialektatlas von Ägypten, Wiesbaden.
Blanc 1981: H. Blanc, “Egyptian Arabic in the Seventeenth Century. Notes on the Judeo- Arabic Passages of Darxe Noʾam (Venice, 1697),” in: Studies in Judaism and Islam. Presented to Shelomo Dov Goitein on the Occasion of his Eightieth Birthday by his Students, Colleagues, and Friends, Jerusalem: 185–202.
- Search Google Scholar
- Export Citation
)| false Blanc : 1981 , H. Blanc “Egyptian Arabic in the Seventeenth Century. Notes on the Judeo- Arabic Passages of Darxe Noʾam (Venice, 1697),”in: Studies in Judaism and Islam. Presented to Shelomo Dov Goitein on the Occasion of his Eightieth Birthday by his Students, Colleagues, and Friends, Jerusalem: 185– 202.
Diem 1985: W. Diem, “Die monophthongisierung der Diphthonge ay und aw im frühen Neuarabisch,” Zeitschrift für Arabische Linguistik 14: 76–78.
Dietz 1999: K. Dietz, “Die gotischen Lehnwörter mit au im Altprovenzalischen und die Rekonstruktion des gotischen Lautsystems,” Sprachwissenschaft 24: 127–156.
Lévi-Provençal et al. 2012: E. Lévi-Provençal, J. D. Latham, L. Torres Balbás and G. S. Colin, “al-Andalus,” in: Encyclopaedia of Islam, 2nd ed., eds. P. Bearman, Th. Bianquis, C. E. Bosworth, E. van Donzel and W. P. Heinrichs, http://dx.doi.org/10.1163/1573–3912_islam_COM_0054.
- Search Google Scholar
- Export Citation
)| false Lévi-Provençal et al. : 2012 , , E. Lévi-Provençal , J. D. Latham and L. Torres Balbás G. S. Colin “al-Andalus,”in: Encyclopaedia of Islam, 2nd ed., eds. , , P. Bearman , Th. Bianquis , C. E. Bosworth and E. van Donzel W. P. Heinrichs . http://dx.doi.org/10.1163/1573–3912_islam_COM_0054
Fischer 1967: W. Fischer, “Silbenstruktur und Vokalismus im Arabischen,” Zeitschrift der Deutschen Morgenländischen Gesellschaft 117: 30–77.
Iványi 2012: T. Iványi, “Diphthongs,” in: Encyclopedia of Arabic Language and Linguistics, ed. Lutz Edzard and Rudolf de Jong, http://dx.doi.org/10.1163/1570–6699_eall_EALL_COM_0091.
Jomier 2012: J. Jomier, “al-Fusṭāṭ,” in: Encyclopaedia of Islam, 2nd ed., eds. P. Bearman, Th. Bianquis, C. E. Bosworth, E. van Donzel and W. P. Heinrichs, http://dx.doi.org/10.1163/1573-3912_islam_SIM_2409.
Melvinger 2012: A. Melvinger, “al-Madjūs,” in: Encyclopaedia of Islam, 2nd ed., eds. P. Bearman, Th. Bianquis, C. E. Bosworth, E. van Donzel and W. P. Heinrichs, http://dx.doi.org/10.1163/1573-3912_islam_SIM_4752.
Nielsen 1981: H. Nielsen, Old English and the Continental Germanic Languages. A Survey of Morphological and Phonological Interrelations, Innsbruck.
Norlin 1985: K. Norlin, “Acoustic Analysis of Vowels and Diphthongs in Cairo Arabic,” Studies in African Linguistics 16: 238–244.
Rosenbaum 2003: H. Rosenbaum, “Another Egyptian Dialect: Spoken Jewish Arabic in Egypt in the Twentieth Century,” in: AIDA 5th Conference Proceedings, Cádiz, September 2002, eds. Ignacio Ferrando and Juan José Sánchez Sandoval, Cádiz: 545–560.
Woidich 1991: M. Woidich, “Short /a/ in Cairo Arabic Morphophonology,” in: Semitic Studies. In Honor of Wolf Leslau, Vol. II: 1632–1651.
Woidich 1997: M. Woidich, “Egyptian Arabic and Dialect Contact in Historical Perspective,” in: Humanism, Culture and Language in the Near East. Studies in Honor of Georg Krotkoff: 185–197.
I owe this suggestion to dr. Marlies Philippa, who has also been very helpful in finding literature about Germanic and in translating Scandinavian works.
Radical consonants are the 3 consonants of the root of Arabic words.
Fleisch 1961, 147: ṯawbbakrin < ṯawbu bakrin ‘the clothing of Bakr’ and ğaybbakrin < ğaybu Bakr ‘the split (in the vest) of Bakr’.
Hijra is the moving away of Mohammed and his followers from Mecca to Medina in 622 after Christ.
Blanc (1981, 195) mentions transcriptions by Forskål (Niebuhr, 1772) and Maillet (1735).
-ēn is a common dual ending.