2.1 The geographic context

The area selected for the toponymic study is located within the UTM zone 34, between coordinates 561000 and 626000 E, and with a latitude between 3616000 N and the coastline, which encompasses a total area of about 1,670 km². In linear terms, the territory described has a length of 65 km and a width that varies depending on the distance to the coastline, with a maximum of 30 km from the capes of ʿĀmir and Raʾs al-Hilāl, and a minimum of 22.5 km from the area of Latrun. This polygon has been defined exclusively by the use of cartographic criteria.

The area includes the following towns (marked on the map in Figure 2 by striped polygons): al-Bayḍāʾ (on the map: al-Bayda’), Šaḥāt (Shahat), which is the old Cyrene, and al-Qubba, together with smaller places such as Šūsa (Susa), al-Abraq (al-Abrak), Lamlūka (Lamluda), Qamāda (Qamadah), al-Qayqāb and Qaryat al-Faydiyya (on the map: Faydiya). The map shows the airport at al-Abraq, which is also indicated by a striped polygon.

Figure 2.

The study area of toponyms with inhabited centres.

From the orographic point of view, the portion of territory studied is divided into three terraces that go from north to south from the coast to inland areas, separated by escarpments with steep slopes. The slope map shown in Figure 3 clearly shows these three units and the steepness of the slopes.

Figure 3.

A slope map of the area.

The lower terrace is a very narrow coastal strip, separated from the middle one by very steep escarpments that close it towards the south. There is a twisted coastline where cliffs are the major feature. This has meant that the number of coastal settlements is very small and that port activity has been negligible. The exception is the port of Sūsa, ancient Apollonia, which, with the arrival of Christianity, changed its name to Sozusa. As the port of the polis of Cyrene, Apollonia had been one of the most important centres in the region ever since the Hellenistic period (Stucchi Reference Stucchi1975, 577–79; Ward-Perkins and Goodchild Reference Ward-Perkins, Goodchild and Reynolds2003, 35–113).

The middle terrace is located above the previous one at a height of between 400 m above sea level on the southern side and 75 m above sea level on the northern, coastal side. It is crossed transversally by numerous ravines that have been produced by the erosion of torrential water courses, the wadis, in their outflow to the sea. These ravines have a highly branched appearance which contribute to a multitude of generally small areas framed by deep ravines of steep slopes, accessible only from one side, usually the south side, which allows access to the terrace. These features almost certainly account for the fact that currently there are no towns in the area. There are two very distinct areas: the western area, consisting of an open plain with wide gaps between the ravines, and the eastern area, where the space narrows considerably, pressing in on the ravines and creating smaller areas with even steeper cuts.

The southernmost top terrace is over 550 m above sea level. To the north, it is closed by a slope that falls almost vertically onto the middle terrace (see the slope map in Figure 3). It is made up of a large plain, which rises gently and gradually southwards without any steep slopes or rugged ravines. Only on the east and west flanks do ravines appear, which are very pronounced in the western area and gentler in the east. It is on this plain that the main human settlements are located today, as shown in the map in Figure 3, represented as striped polygons.

2.2 The mapping of the toponyms

All three sources used in this study to obtain toponyms are of recent creation (as the purpose is not to identify the sites mentioned in historical written sources): colonial era maps, online databases and published archaeological studies.

With regard to the cartography from the colonial period, maps were used that were made by the Italian army in 1931–36 and then reproduced in 1942 by the US Army at 1:50,000 scale,Footnote ⁶ where a high number of toponyms are collected of places large and small, many of which have now disappeared. As for the online databases, the GeonamesFootnote ⁷ cartographic database has been extremely useful. The information from these sources was later compared with the main archaeological syntheses of the area (Stucchi Reference Stucchi1975; Ward-Perkins and Goodchild Reference Ward-Perkins, Goodchild and Reynolds2003), where several toponyms are collected and linked to archaeological remains.Footnote ⁸

The mapping of the locations was conducted in two stages.Footnote ⁹ First, the toponyms were placed on the geographic locations that were shown in the different cartographic sources. Second, this georeferencing was corrected wherever possible by consulting archaeological studies – in particular the syntheses made by Stucchi (Reference Stucchi1975) and Ward-Perkins and Goodchild (Reference Ward-Perkins, Goodchild and Reynolds2003)Footnote ¹⁰ – and by visual inspection of a high-resolution satellite photograph accessed through Google Earth, which allowed for the locating or testing of any material evidence that may correspond to each toponym; finally, when conditions allowed, the location was visited for exploration on the ground.Footnote ¹¹

The satellite photograph was only taken into account for those traces where there were perceptible structures carved into the rock or where indications of the existence of very distinctive buildings could be seen, such as those with square or rectangular floor plans, with canted buttresses often associated with a perimeter moat, identified as basilicas during the Byzantine period by Stucchi (Reference Stucchi1975, 430–34).Footnote ¹² Once the material evidence was found, the correspondence between the remains and the toponym was established only when the following circumstances occurred simultaneously:

1. 1. the remains identified were located no farther than 500 m from the location shown in the cartographic sources consulted;

2. 2. between the geographic point and the material remains there were no substantial natural barriers, particularly ravines, which could be described as being insurmountable due to their steep slope.

In cases where there was no such match, which were the majority, the toponym was assigned to the location attributed to it in the cartographic sources consulted, pending a future archaeological survey that would enable the remains to be identified on the ground.

3.1 The first term of the toponyms: a common element semantically associated with the idea of settlement

This study is limited to those toponyms composed of several terms, the first of which refers to the existence of a human settlement by the use of the words qa ṣr, referring to a ‘fortress’, and sīra and qarya, both terms that can be translated as ‘hamlet’.

The first group is constituted by the toponyms with the form qa ṣr (قصر) (see map in Appendix 1). The generic meaning of this word refers to a castle or a palace. As an integral part of toponyms, the term qa ṣr is documented throughout the Arab world, from al-Andalus to Iraq, whether in the singular (qaṣr), plural (quṣūr), dual (qaṣrayn) or diminutive (quṣayr) forms.

From the point of view of the material records, there have been numerous studies in recent decades about these qu ṣūr, of particular interest those conducted in the area of Syria and Jordan.Footnote ¹³ In this vast region, it has been concluded that these quṣūr are rural establishments, which have a certain poise or a monumental structure and act as a centre for agricultural farms, something that Grabar (Reference Grabar1973, 139–78) drew attention to in the 1970s. In terms of chronology, they are usually dated as being from the period of government of the Marwānid branch of the Umayyad dynasty, between 685 and 750, although there are cases of foundations from both of the ʿAbbasid periods and of quṣūr built on top of earlier constructions from the Roman period.

However, studies on Libya refer to a different situation which is unique in the Islamic world. The first typology of qu ṣūr is due to Goodchild (Reference Goodchild1950, 35–37), who attributed to these buildings a different chronology depending on the construction technique, covering the long period between the beginning of the Roman occupation and the early centuries of the Islamic period. The team led by Barker, who during the 1980s and 1990s worked intensively on the survey and location of settlements in a vast region of the Tripolitania area, went even further. According to Barker (Reference Barker and Mattingly1996, 112–33 and 326–32), the quṣūr of this region were foundations from the Roman period associated with the processing of agricultural products, as seen by the existence of oil presses and storage items, which had already been out of use years before the Islamic Arab occupation. In this way, the original meaning of the Arabic word qaṣr became completely diluted, as it was used to refer to a settlement of Roman chronology, that is, one different from its own.Footnote ¹⁴

However, the purpose of this study does not lie in the material records, but in the toponymic records, which ultimately refer to a word of Arabic origin. In the territory under analysis, the number of toponyms whose first element is qa ṣr is overwhelming, with a total of 89 names that correspond to 73 sites.Footnote ¹⁵ Their distribution is very heterogeneous and it is not possible to establish a pattern on the map today. They appear on the plains of the upper terrace (18 cases); on the ravines that constitute the border of this terrace on the west (10 cases); on the slope that separates the middle terrace (14 cases); at the foot of the referred slope (5 cases); and in the gaps formed between the ravines of the middle terrace (22 cases). Strikingly, there is only a single record of this kind in the coastal terrace.

The second group is that constituted by toponyms preceded by the form sīra (سيرة) (see map in Appendix 2). In Cyrenaica, a vast number of toponyms have been documented with the form sīra, which is, in turn, virtually non-existent outside the region being documented. Although its first meaning is ‘path’, the abundance and random distribution of the sites with this element as the first term seems to refer to a local word that designated small rural settlements linked to agricultural activities. A similar situation exists in Tunisia and Tripolitania with the word hanšir, with the word ra ḥl in Sicily (Metcalfe Reference Metcalfe2003) and the eastern coast of the Iberian Peninsula, and with the word ḍayʿa – in Spanish aldea, in Portuguese aldeia (hamlet) – in the interior and eastern coast of the Iberian Peninsula. The frequent repetition of all these local words in the toponymic records contrasts vividly with their absence from the written records – that is, in Arab chronicles – which refer to such sites generically with words such as qarya. Everything seems to suggest that it is confined to local idioms restricted to the spoken language and fossilised in the toponymy.Footnote ¹⁶

In the area studied, a total of 43 sites with the name sīra were recorded, which corresponded to 49 toponyms. It should be noted that, on five occasions, places were recorded whose toponyms were formed interchangeably by sīra or by qa ṣr (S-14 = Q-22, S-23 = Q-48, S-24 = Q-50, S-34 = Q-67, S-36 = Q-68, S-41 = Q-71),Footnote ¹⁷ and Qaṣr al-Suwayra (Q-35) was documented, where the second term, ‘Suwayra’, turns out to be the diminutive of sīra. Their distribution throughout the territory is very heterogeneous, although the map shows a significant gap in the central section of the plain of the upper terrace, with a tendency to be at the sharp edge of the terrace or in the interstices between ravines of the lower terrace. Moreover, there is no single toponym preceded by the form sīra in the coastal strip.

Finally, the last group is formed of toponyms preceded by the form qarya (قرية) (see map in Appendix 3). The word qarya generically designates rural settlements, that is, villages or hamlets, as opposed to the madīna or city, and its presence in toponyms is common and very abundant throughout the Arab world. In the portion of the territory of this study, however, its occurrence is minimal, probably displaced by the local word sīra, and is limited to three cases. One is currently a town: Qaryat al-Faydiyya.

3.2 The second term: the individualisation of the common element

Compared with the first term discussed, which has a generic nature, the second term serves to identify that place in relation to those around it. With few exceptions, due to the use of common nouns, it is unique to each place and constitutes its hallmark. In the toponyms analysed, this second term can be an adjective in agreement with the noun, in which case both terms seem to be preceded by the article al- and agree in gender and number, or, most often, they may refer to an element that is associated with that place, usually expressed by means of a noun or a nominalised adjective. This element is sometimes expressed by a common noun: a crop, a landform, and so on. But, in a very significant part of the records, as will be seen, the second term consists of a proper noun with which that settlement is associated. In these cases, where the element designated by the second term of the toponym refers to the name or nickname (laqab) of an individual, it should be understood to most likely be the eponym of a lineage with which the site in question is associated.

The study analysed a total of 145 toponyms that corresponded to 118 sites.Footnote ¹⁸ They were distributed as follows: 89 toponyms for the 73 sites with the form qa ṣr; 53 for the 42 sites with the form sīra; and three toponyms for another three sites with the form qarya. Many of the toponyms collected are enigmatic in their interpretation as a result of the limitations inherent in toponymic records. Nevertheless, the number of toponyms subject to interpretation was 59 with the word qaṣr, 40 with the word sīra and three with the word qarya, that is, 70 per cent of the total, which provides a highly significant sample of the overall set.

The toponyms containing the initial form qaṣr are the most abundant. Qaṣr appears associated with words related to the agricultural world, particularly to the harvest, in Qaṣr Qaṭṭufa (Q-19); to a widespread local agricultural product, the olive, in Qaṣr al-Zaytūn (Q-9); or to a product associated with livestock and straw in Qaṣr al-Tibn (Q-55). The form qaṣr also appears to be associated with a water resource, a pond or pool in Qaṣr al-Birak or al-Burūk (Q-39), and with a dry stream bed in Qaṣr al-Baṭḥāʾ (Q-33). The toponym Qaṣr al-Ǧubayil (Q-29) refers to a landform, specifically, a hill.

Toponyms have also been documented that associate qa ṣr with words designating animals, such as the swift in Qaṣr al-Samīm (Q-21) or wolves in Qaṣr al-Ḏiʾāb (Q-67). These toponyms, however, should not be understood as references to animals, but as the laqab or ‘nickname’ of a lineage. This can be concluded for the toponym Sīrat Umm al-Ḏiʾāb (S-34), another name given to the same place, where the word ‘Ḏiʾāb’ appears associated with the term ‘Umm’ (mother), thus forming the kunya, in this case the matronymic, which refers to one of the elements that make up the name of a people in Arabic (Schimmel Reference Schimmel1989, 1–13).

On several occasions, terms associated with settlements are recorded, as in the case of Qaṣr al-Madīna, or ‘of the city’ (Q-64); Qaṣr al-Quraya, the diminutive form of qarya, that is, ‘hamlet’ (Q-73); and Qaṣr al-Suwayra (Q-35), the diminutive form of sīra. There is also a qaṣr that appears to be associated with an artificial cave, Qaṣr al-Ḥuwayfir, the diminutive of ḥufra (Q-4), and two other toponyms related to surveillance tasks: Qaṣr al-Nūriyya (Q-43), a word related to fire, which could be linked to a signal tower or minaret, and Qaṣr al-Murayqib (Q-8), again a diminutive which means a watching place.

Especially abundant are toponyms generically associated with ruins or the remains of buildings, which are included in the toponyms Qaṣr al-Ḫirba (Q-27), its diminutive form Qaṣr al-Ḫurayba (Q-2, Q-28 and Q-38) and Qaṣr al-Ṭalūla (Q-24). A Qaṣr al-Ṣūmaʿa, or ‘minaret’, is also documented (Q-62), and it is particularly important because the toponym seems to suggest the existence of a mosque.

Most often, the term qa ṣr appears in connection with a person's name or, more likely, a lineage traced through their eponym. Particularly significant are the names that are designated by the kunya, that is, the name preceded by the forms Bū (i.e. Abū, meaning ‘father’) and Umm (mother). As already noted, on several occasions the existence of the kunya has served to signify that the common noun or adjective that accompanies the Bū / Umm actually refers to a lineage by means of the laqab of its eponym. As for the toponyms formed by the term Bū, these were documented a total of 13 times, designating nine places (Q-3, Q-7, Q-13, Q-30, Q-40, Q-42, Q-53, Q-54 and Q-71). The feminine form, Umm, appears five times to designate five locations (Q-11, Q-16, Q-56, Q-60 and Q-61). In other cases, the name of the person is reflected by the nisba, or ‘place of origin’, of the subject in question (Q-1, Q-48 and Q-56).

On several occasions the individual's name is represented by a common noun or an adjective used as a laqab. Thus, in the case of common nouns, a particularly significant toponym was documented where the word qa ṣr appears associated with the word al-Raqīq, ‘the slave’ (Q-59), revealing the status that the eponym of the lineage had at some point. As for the names of animals, the presence of these nouns in the kunya, as part of another toponym, shows that it was a laqab. A further case was also documented where a qaṣr was associated with the lineage of the Qurayẓa (Q-73), Jews who had lived in the vicinity of Yaṯrīb, now Medina, and were defeated by the Muslims in the early years of Islam.Footnote ¹⁹

The presence of certain adjectives such as al- Ṭawīl, the ‘Tall One’ (Q-63), can also be attributed to the laqab. The inclusion in this section of the toponym Qaṣr al-Bayḍāʾ, the ‘White One’ (Q-5) is dubious; while nothing prevents it from being interpreted as the laqab of a woman, along the lines of the female kunya, it should not be forgotten that there was a nearby site of this name, which was very common in the Arab world as an epithet for towns.

Finally, there are some toponyms that seem to refer to non-Arab words. A Qaṣr Nūṭa (Q-25) was documented that is very likely to conceal the Greek word nautai, ‘sailors’, and is notably the only toponym with the form qaṣr located on the coast.Footnote ²⁰ A similar case is that of the toponym Qaṣr Batruh (Q-20), under which a form baṭrīq may be hidden, derived in turn from the word patriarcha.Footnote ²¹

The toponyms containing the initial form sīra are also abundant. Starting with the adjectives that agree with the noun sīra, a Sīra Šabūba (S-13), or ‘beautiful’, was documented. There is also a descriptive toponym that refers directly to the name of the sīra: this is the Sīra Ṭayyibat al-Ism (S-40) or ‘of the Beautiful Name’, although paradoxically that name has not reached us.

Regarding common nouns, as in the previous section, some words were recorded which were related to agricultural production. Thus, two vineyards (karm) appeared in Sīrat Karm al-Qada (S-6) and Sīrat Karm al-Ḥisān (S-9), but also wild herbs such as thyme in Sīrat al-Ṣaʿtar (S-38) and hawthorn in Sīrat al-Sidr (S-22). Regarding water features, a Sīrat al-Ḥawḍ (S-26) was recorded, which refers to a pool or pond. Another toponym, Sīrat al-Ṣafṣaf (S-13), was also identified, which refers to the plain. Halfway between geographic relief and communication routes, there is a Sīrat al-Bāb (S-28) or the ‘Mountain Pass’, located at the entrance to the ravine along which the road that connected the city of Limnias with the coast once ran (Johnson Reference Johnson1973, 116–17).  In relation to human settlements, only one toponym was registered, although redundant in nature, as it repeats the word sīra but in its diminutive form:  namely, Sīrat al-Suwayra (S-3).

As with the qa ṣr-type toponyms, there are abundant place names that refer to the kunya, with a total of eight names in the masculine – that is, with the form Bū – corresponding to six locations (S-4, S-7, S-8, S-15, S-16 and S-20), and five in its feminine form, Umm, corresponding to another five (S-1, S-5, S-34, S-39 and S-41). Precisely one of the latter has the form Sīrat Umm al-Ḏiʾāb, ‘Mother of the Wolves’, where the animal's name obviously alludes to the laqab of a lineage. The same could be assumed in the other cases of references to animals, such as Sīrat Saqr, ‘Falcon’ (S-29), Sīrat Kalb, ‘Dog’Footnote ²² (S-32), and Sīrat Ġazīla, ‘Gazelle’ (S-27). One place has two different names, both with nisba form: Sīrat al-ʿAwsaǧī and Sīrat al-ʿAwsazī (S-42).

There are other toponyms whose second term has the seamless look of an Arabic word but whose meaning is somewhat strange for the creation of a toponym, which suggests a semantic shift caused by the Arabisation of a non-Arabic term. This is the case for Sīrat al-Bišāra, the ‘Good News’ (S-33). We do not know the term underlying the Arabic word and, therefore, it remains a mere hypothesis.Footnote ²³ Other toponyms, however, seem to have underlying Greek or Latin names in their current form, such as Sīrat Martuniya (S-2), Sīrat Barqū (S-17) and Sīrat Yara (S-24), whose etymology is also problematic to establish.

Finally, the toponyms containing the initial form qarya are only three: Qaryat al-Faydiyya (QY-1), which refers to the nisba of a woman; Qaryat Yūnus (QY-2), which refers to the ism or the name of a man; and Qaryat Umm Sulaymān (QY-3), which refers to a matronymic, the ‘Mother of the Sulaymān’ referred to in the name.

4.1 Settlements located on the centre of the plain in the upper terrace

The upper terrace is characterised by gentle slopes and plenty of meanders formed by wadis with a wandering course that provide moisture for the soil, making it more fertile, as can still be seen today. The settlements identified with toponyms located in this plain are, therefore, the most privileged in terms of soil quality and are found in places surrounded by farmland, sometimes occupying small promontories, a characteristic already noted by Stucchi (Reference Stucchi1975, 433–34). Another feature of these sites is accessibility, favoured by the gentle slopes but also by the road passage which, according to Itinerarium Antonini (Reference Parthey and Pinder1848, 68 and 70), crossed the territory in an east–west direction joining Darnis (Derna) with Tolmeta (Ptolomaide) through the inland route by way of Cyrene (Purcaro Pagano Reference Purcaro Pagano1976, 292). All these observations would lead one to think that the sites were closely related to the production of surplus products for trade in the Mediterranean. We must not forget that some of the toponyms located here preserve their Greek roots, despite their apparent Arabisation (Stucchi Reference Stucchi1975).Footnote ²⁴

These sites occupy central places in the context of the surrounding production areas, where the cultivation of vines and olive trees must have prevailed, forming the basis of their livelihood (Wilson Reference Wilson2004, 148–49). This very accessibility, which is a positive feature for agricultural spaces, becomes problematic in terms of the sites’ vulnerability. This accounts for the enclosing walls found around some of the sites in the area.

4.2 Settlements located on the edge of the upper terrace and on the more open plains of the middle terrace

Unlike the previous ones, these settlements occupy areas on the outskirts of the best production areas. They are situated on the edges of the upper terrace or on the western section of the middle terrace, where the plain is more spacious. The communication routes are located at a certain distance from the Derna–Barqa road, currently Marǧ, which may have been accessed through branch roads. They are likewise found in the vicinity of the road that linked Cyrene with Apollonia, currently Sūsa, which also must have had significant freight traffic.

There was easy access from these settlements to the wide, open spaces on the plain of the upper terrace so, despite occupying marginal spaces, presumably they made their livelihood from these fields. These settlements are substantially better placed for defence purposes than the previous ones, as they are often protected by significant natural defences on various sides.

4.3 Settlements located on the edge of the ravines in the interstices of the middle terrace

A small – but not negligible – group of settlements is located in the interstices of the middle terrace, occupying land that is totally marginal from the standpoint of productivity and accessibility. In relation to their surroundings, these settlements are in the centre of small production spaces. They are very stony and confined by remarkable natural barriers, particularly by deep ravines that limit access to these parts to one of their sides, often through narrow, steep and intricate passes.

The remote access paths to these isolated places suggest that they rarely produced surplus commodities that could be used for trade or even exchange. On the contrary, it seems that these were self-sufficient settlements, endowed with significant natural defences, where all production was destined for their own consumption.