Introduction

The emergence of Islam in Mecca and Medina and its revelation in Arabic seem to be, prima facie, a clear indication for the emplacement of early Islam in the Western 166 Arabian Peninsula. However, so-called “revisionist” and “neo-revisionist”¹ scholarship— which, according to its own claims, contests parts of Orientalist scholarship as well as most parts of the Islamic tradition—has started to question the location as well as the linguistic unambiguity of early Islam with an ever-growing verve in the last two decades. Revisionist scholarship has especially cast doubt upon Mecca and Medina as the birthplaces of Islam.

Mainstream and revisionist scholarship attributes much attention to the primordial places of Islam, either in order to uphold the traditional chronology and emplacement of early Islam or in order to deconstruct and reconstruct the historical centre of Islam. My objection to both approaches is that a limited understanding for the historical production of space is at work here, so that neither mainstream nor revisionist scholarship can adequately treat the intricate questions that they want to solve. That dominant scholarly approaches either take for granted or relocate the birthplace of Islam leads to a confusion between different spatial dimensions, especially between the physical and symbolic ones. It also blurs the boundaries between Qurʾānic genesis and exegesis and takes chronology for causality.

The question of what we actually know about the places in which Islam emerged in Late Antiquity is certainly worth a discussion. Revisiting Islamic origins with new methods is “an important desideratum,” as scholar of religions studies Aaron Hughes (2017, 883) put it; yet this is certainly not so because the origins of Islam are different from the beginnings of other religious traditions. Instead, “there is some discrepancy between sources and tradition, as there are in all religions” (ibid., 872); the origins of Islam “are clouded in mystery, and are about human ingenuity and worldmaking in the midst of rapid change” (ibid., 869). What is different with Islam, however, are “the many political and ideological uses to which the discourse of Islamic origins have been put over the years” (ibid., 871). Because of this effect, revisionist discussions partly verge on Islamophobic sentiments, especially after 9/11, when the “Syro-Aramaic reading” of the Qurʾān hit the headlines by poking fun at martyrs, who will find, if at all, grapes, not virgins, in paradise (Luxenberg 2004; for a critique Wild 2010; Saleh 2010; Sinai 2012).

Having said this, I understand my critique of the debate about the emergence of Islam as a contribution that questions the heterogeneous scholarly research which implicitly aims to explain what Islam ‘really’ is, or is not, by looking at its possibly "dark 167

origins" (Ohlig/Puin 2005). I do not concur with the diagnosis that Qurʾānic studies are, because of revisionist activities, in a state of “disarray” (Donner 2008, 29), since such “an assessment is best understood as a rhetorical appeal to the writer’s own approach as compared to those of rival theories,” as Andrew Rippin (2013, 60) has succinctly noted. I also cannot join in the enthusiasm about “the large-scale interest of the media that the Qurʾan’s origin and interpretation have solicited during the last decade or so” (Sinai/Neuwirth 2010, 1). Instead, I understand the historiography of early Islam not only as a complex, but a “wicked problem”² that is characterized by “intense disagreement between fragmented stakeholders, multiple and often conflicting objectives, as well as high levels of uncertainty, variability, and risk” (Lake 2014, 77). By participating in this debate, I question the self-affirming knowledge produced along institutional boundaries and ossified positions in Islamic tradition, mainstream scholarship, and revisionism. Based on the assumption that the underlying questions are not amenable to final resolutions, I argue for “supplanting initial divisive certainties with ‘perplexities’” (ibid., 80). From a “sympathetic understanding” (ibid., 84) for different perspectives, I have invited these into a critical dialogue on the following pages. From an epistemological point of view, my critique focuses on three overlapping fields of knowledge production. Firstly, I am sceptical about the results of scholarly “parallelomania” (Sandmel 1962) that incessantly discloses similarities between the three monotheistic religions for paternalistic or defamatory reasons, although scholarship on Late Antiquity is certainly aware of these dangers (e.g. Koloska 2016, 431). Secondly, my critique follows from the impression that a good deal of scholarship in Qurʾānic studies confuses physical and symbolic space and is trapped either in age-old tropes on Islam or in a sort of back projection of later knowledge into the text. Thirdly, I question the general meaningfulness and importance that is acceded to scholarship on early Islam against the background of current popular debates on Islam.

In order to meet this objective, I firstly give an overview of the spatial knowledge that we have with regard to early Islam, especially with a focus on the absence of topographical terms in the Qurʾān. Then, I turn to the main arguments and counter-arguments that speak in favour of or against a relocation of the birthplace of Islam. Next, I highlight the shortcomings in this debate. On the revisionist side, I focus on the lack of a new comprehensive chronology, the philological bias (or rather a biased philology), and the reductionism that focuses on Jewish-Christian text layers in the Qurʾān. On the mainstream side, I pinpoint, pars pro toto, some central findings from Qurʾānic studies in order to show that some researchers’ emplacement of the Qurʾān blurs the historical and symbolical dimensions of space. Finally, I propose an alternative 168 topological reading that tries to elucidate the production of a multi-religious landscape as well as a Muslim space in the Ḥijāz as a lasting process that cannot be reduced to the first decades of the history of Islam. My intention is to establish multi-religiosity as a central category to explain the process that led to the emergence of Islam in the first centuries of the history of Islam. In a diverse environment, practices of “un/doing differences” (Hirschauer 2014) were present from the very beginning and included different practical and symbolic actions, such as instances of sharing, the creation of ambivalence, and processes of purification.

At the Origins of Islam

It is a well-established fact that the Qurʾān has little concern with its own historical and geographical context (e.g. Reynolds 2010a, 198f.). Because of that, context is given to Qurʾānic verses by the two genres of sīra, biographies of the life of the Prophet Muḥammad, and tafsīr, comments on the Qurʾān, both of which developed approximately 150 to 200 years after Muḥammad’s death in 632. Without this additional literature, it is difficult to understand the Qurʾān on its own. The Qurʾān mentions Mecca only once explicitly (Q 48:24), but does not identify it as a city, as it merely hints, rather vaguely, at “the heart of Mecca” (baṭn al-Makka)³. In another verse (Q 3:96), the commentators identify the term “Bakka”⁴ with Mecca. Similarly, the Qurʾānic use of al-madīna (Q 9:101, 9:120, 33:60, and 63:8) “is too general to assume that it is a proper name” (Reynolds 2010a, 199n713), while Yathrib is never mentioned—a lack that is comparable to many other locations in which Biblical or Qurʾānic events are assumed to have taken place. Reynolds (2010a, 199) mentions the lack of Uḥūd, Ṭāʾif, Arabia, Egypt, Yemen, Persia, and the Red Sea. Muḥammad’s hijra, his move from Mecca to Yathrib/Medina, which represents a crucial event for the early Muslim community and is held to have taken place in 622 CE, is also not mentioned in the Qurʾān. Jerusalem, which was later called the third holy city of Islam, is also missing in the Qurʾān, although the “further mosque” mentioned in verse 17:1⁵ is traditionally understood as describing the Prophet’s “night journey” from Mecca to Jerusalem. Yet the designations for Mecca (“the holy mosque,” 169 al-masjid al-ḥaram) and Jerusalem (“the further mosque,” al-masjid al-aqṣā) are open to interpretation, if one puts aside sīra and tafsīr literature for a moment. In other cases, the reference point of Qurʾānic expressions is also unclear. The word kaʿba, for example, is only mentioned two times (Q 5:95, 5:97), but not explicitly in connection with the ḥajj (pilgrimage) or the qibla (direction of prayer). Muslim commentators refer expressions like bayt (house), maqām Ibrāhīm (site of Ibrahim), and al-masjid al-ḥaram to the place of worship in Mecca. However, with regard to the pilgrimage, the Qurʾān simply directs people to “the house” (Q 3:97). The qibla is advised as the direction of prayers with the order to “turn your face to al-Masjid al-Ḥaram” (Q 2:144), although at the time of the change of the qibla from Jerusalem to Mecca—according to Islamic tradition, in the year 623/24—the later mosque was still a place of pagan worship. The term miḥrāb, which is later used to describe the concave niche in the wall of a mosque, which indicates the prayer direction (to Mecca), means something else (“temple”) in the Qurʾān and early Umayyad texts, has an unclear etymology, and appears as a prayer niche in the early eighth century, according to Muslim historians (Khoury 1998).

A similar Qurʾānic silence holds true for the main protagonists of early Islam who are attested in sīra a ccounts, l ike " the P rophet’s w ives K hadīja a nd ʿĀʾisha, (…) h is daughter Fāṭima, his uncle Abū Ṭālib, his cousin ʿAlī, or his companions Abū Bakr, ʿUmar, and ʿUthmān" (Reynolds 2010a, 199). The Byzantines and the ruling Meccan tribe of the Quraysh are also only mentioned once in passing, without giving any details (ibid.). On top of that, the Qurʾān gives next to no biographical details about Muḥammad. One can even argue that “the Qurʾān never identifies the speaker or the intended audience” (Reynolds 2010a, 15) of its message, even when it directly addresses this speaker or his community and uses different names, epithets, or titles for him. While the Qurʾān abounds with references to a messenger (rasūl) or prophet (nabī), Muḥammad only appears four times (Q 3:144, 33:40, 47:2, and 48:29), while Abraham, Moses, and Jesus are named 69, 136, and 25 times, respectively; even Mary appears 34 times and 70 verses allude to her, more than in the Gospels. Against this backdrop, the Islamic tradition—as expounded, for example, by Ibn al-Jawzī (d. 1200)—suggests that the Prophet had up to 23 (or even more) names, titles, or epithets, among them muḥammad (“the praised-one”), aḥmad (“the most-praised one”), and al-amīn (“the reliable”) (Reynolds 2010a, 193). However, it remains unclear whether muḥammad is a proper name or a title. The four times a muḥammad is mentioned in the Qurʾān, the verses refer to his vocation as the messenger of God, which could also refer to other messengers, like Jesus, whereas when “a messenger to come after me” is announced in the words of Jesus (Q 61:6), he is called Aḥmad.

The historical, biographical, and geographical context of the Qurʾān is mainly added by later hagiographic and exegetical literature. In the Islamic tradition, the Qurʾān is so 170 densely inter-connected with sīra and tafsīr works that each genre helps to explain the lacunae of the other. In the traditional process of interpretation, Qurʾānic verses, which were revealed to the Prophet step by step, are not only connected with the events of his biography, but their meaning is also explained with reference to these events. In this sense, the Qurʾān acts as proof of Muḥammad’s life, while Muḥammad’s deeds are a reflection of the Qurʾān. Insofar as it is difficult to distinguish religious legends from historical facts, not only John Burton’s (1993, 271) remark that “exegesis aspiring to become history, [sic] gave us sīra” is worthy of consideration. Beyond this, the more fundamental question is whether sīra and tafsīr literature works by reading meaning into Qurʾānic verses rather than by reading meaning out of them. The fact that the Qurʾān is closely connected with reports about Muḥammad’s life, composed much later, gives the impression that this kind of understanding of the Qurʾānic message comes closer to eisegesis than exegesis (Bangert 2016, 10).

The eldest testimonies of Qurʾānic verses were identified in the Dome of Rock in 692 CE, on Arabic coins from the time of caliph ʿAbdalmalik (r. 685–705), attributed to 692 and 697/98 CE, and as tombstone inscriptions from Egypt, attributed to 691 CE (Bacharach/Anwar 2012). The collection and redaction of the Qurʾān under the caliph ʿUthmān b. ʿAffān (r. 644–656) resulted in the so-called rasm ʿUthmānī in the mid-seventh century, according to Muslim and mainstream academic scholarship. Under the reign of caliph ʿAbdalmalik, an orthographic reform was implemented which introduced diacritical dots for consonant homographs (Würsch 2013, 34f.). The introduction of vowels for words with different spelling possibilities happened by the end of eighth century. While the Islamic tradition knew different readings as well as some minor textual variants, the printed standard edition by Azhar University, from 1923/24, follows the reading of Kūfa (Ḥafṣ ʿan ʿĀṣim). Yet the underlying manuscripts of the Azhar edition are unknown, and several inconsistencies in the text have been highlighted (Puin 2011). It is also unclear whether the text really reflects the rasm ʿUthmānī. A critical reconstruction of the textus receptus, based on different existing Qurʾān manuscripts, is still missing (Gilliot 2006, 52), although the Berlin project Corpus Coranicum aims at compiling a historical-critical text edition of the Qurʾān (Würsch 2013, 28). Among Shīʿī scholars, the idea of a “falsification of the Qurʾān” (taḥrīf al-Qurʾān) at the hands of Sunnī redactors who supposedly aimed to erase or belittle the role of ʿAlī b. Abī Ṭālib (d. 661) was widespread up to the tenth century CE; although it subsequently lost its importance, it never totally vanished (Würsch 2013, 34; Brunner 2001). Muslim as well as non-Muslim scholars have occupied themselves with the question of misspellings (taṣḥīf) in the Qurʾānic text. The “dog” in the originally Christian legend of the Seven Sleepers (Qurʾān 18: 9–26) was recently identified as a possible candidate for such a misspelling (Würsch 2013, 35f.; Waldner 2008), since it is not existent in the original 171 legend. Without diacritics, kalbuhum (“their dog”) in verse 18:18 resembles the word kāliʾuhum and could therefore have meant “their guard.” Once fixed in the Qurʾān, the dog of the Seven Sleepers, however, came into his own in popular belief and book paintings and was even counted among the elected creatures who were promised Paradise (Würsch 2013, 36). Squeezing higher sense out of the presence of the dog in the Qurʾānic verse has proved quite challenging intellectually for Muslim interpreters as well as for academic scholars until recently (for some examples, see Koloska 2016, 438).

The Jewish and Christian presence in and around the pre-Islamic Arabian Peninsula is fairly well documented (e.g. Finster 2010; Berger 2016). Yet it is more difficult to ascertain what kind of Christians and Jews lived in the Ḥijāz. Lecker (1985; 1995a; 1995b; 2000; 2016; 2017), sifting through early Arabic literature, tried to figure out whether Muḥammad had monotheistic, Christian, and Jewish relatives, whether Muslims and Jews possibly lived close to each other in Medina, and who the Jewish allies among the Arab tribes were. Jews are regularly mentioned in early Islamic sources, for example in the so-called constitution of Medina; yet the three tribes that were later to be expelled from Medina, according to the tradition, do not appear in the document (Donner 1998, 72f., 227–235; Lecker 2012). According to Islamic tradition, Muḥammad himself ordered an icon with Mary and the Child to be preserved in the Kaʿba after the conquest of Mecca, which suggests that the Kaʿba might have played a role in Christian worship in pre-Islamic and early Islamic times (Rubin 1986, 102; Finster 2010, 83). The name of the Kaʿba itself refers to a cubic architectural body and might have been constructed according to examples of churches in Arabia or the dome at Debra Damo in Ethiopia (Finster 1991; 2010, 76) and seems to have been decorated with paintings and frescos (Finster 2010, 83).

Revisionist Arguments

Against the background of uncertainties and contradictions in the history of early Islam, revisionist and neo-revisionist scholarship—since Crone and Cook (1977) and, with new vigour, since the end of the 1990s—has started to question a growing amount of what these authors call the “assumptions” on which Islamic salvation history as well as traditional Orientalist scholarship is grounded. Challenging mainly the reliability of the earliest available Arabic accounts about the emergence of Islam, revisionist scholars try to disentangle the Qurʾān from the information given by Arab informants and draw, instead, from contemporary non-Arabic sources. In this sense, the scholars separate the Qurʾān and its traditional context from each other and try to reconstruct 172 different relations between the Qurʾān and other—mainly Jewish or Christian—contexts. Revisionist scholars’ doubts about the chronology of early Islam are accompanied by doubts about locations and spaces, traditionally connected with the emergence of Islam. These doubts involve the composition of the Qurʾān, the life of Muḥammad, the role of Mecca, the Arab expansion, and the institutionalization of Islam. While early revisionist scholarship suggested a late composition of the Qurʾān in the ninth century (Wansbrough 1977, Burton 1977, for a critique: Schoeler 2010) and a strong imprint on early Islam by Jewish messianism (Crone/Cook 1977), the neo-revisionist tide has now turned to the possibly Christian antecedents of the Qurʾān and the supposed “Christology” of early Islam (Luxenberg 2004; Kerr 2014a and 2014b; Ohlig 2014). Thus, revisionist scholarship works both ways, suggesting an earlier dating for the Qurʾān and Islamic origins while at the same time proposing a later institutional formation of the Qurʾān and of Islam as we know it today.

Günter Lüling (1993) and Christoph Luxenberg (2004; 2005; 2007; 2008), for example, not only try to re-construct hymnal structures and the meaning of verses, words, and the mysterious letters in the Qurʾān by drawing from non-Arabic vocabulary, but they understand at least parts of the Qurʾān as an originally Christian liturgical text. Revisionist scholars have also suggested that muḥammad should not be understood as a proper name, so that the part of the Islamic creed muḥammadun^un rasūl allāh translates to “Praised be the messenger of God” (Luxenberg 2005; Puin 2014) and could thus refer to Jesus; a prophet of the name Muḥammad might then be seen as literary fiction (Popp 2005). The presumably first Muslim inscriptions in the Dome of the Rock in Jerusalem, with their multiple references to Jesus, could, then, be understood not in the traditional sense as anti-Trinitarian and anti-Christian, but as an anti-Trinitarian yet still Christian writing. A further assertion is that early Umayyads, such as Muʿāwiya b. Abī Sufyān (ruled 661–680 CE), actually were Christians, which is reflected in their use and mintage of coins with Byzantine symbols—a fact revisionists interpret as proof of the existence of a Christology (Popp 2005; for a critique: Heidemann 2007 and 2010).

A recent radiocarbon dating of parchment fragments from old, incomplete Qurʾān manuscripts—one held by Birmingham University, the other stemming from Sanaa, held at the University of Saarland—also added to the discussion about Islamic origins, because the analyses have yielded different results with rather early—or even “too early” (Reynolds 2015)—dates. One folio was dated to 568–645, another to 433–599, when the Prophet, according to Islamic tradition, had not yet received God’s message, and a third one was given a 75 percent likelihood of being older than 646.⁶ All these findings stand in contrast to Muslim historiography, which holds that a committee 173 finished the redaction process for the first official manuscript of the Qurʾān during the caliphate of ʿUthmān (r. 644–656 CE).

Revisionist scholarship has also put Mecca—a name sparsely mentioned in the Qurʾān—as the birthplace of Islam into question. There has been a widespread scholarly consensus for more than 300 years, also corroborated by early Arabic sources, that Mecca was a place of worship in pre-Islamic times already. Claudius Ptolemy, a Greek writer from the second century CE, mentioned several places on the Arabian Peninsula which were later identified with pre-Islamic Mecca, among them a place transcribed as Macoraba. As Morris (2018, 12) recently pointed out, the first scholar who identified Macoraba with ancient Mecca was the Huguenot pastor and Orientalist Samuel Bochart (d. 1667) from Caen, in his Sacred Geography (1646); neither Late Antique nor Arab nor Medieval scholars had ever done so. The background of the concerted efforts of humanists and theologians to find ancient Mecca, which started in the 1530s, was that they tried, with the help of recovered Greek and Latin sources, to reconcile the geography of Mecca with a Biblical genealogy of Islam (ibid., 10). Over time, Bochart’s thesis gained almost universal acceptance, and although it triggered a myriad of etymological explanations,⁷ it turned into a truism in religious studies, holding that Mecca was a centre of worship and trade before Islam, as mentioned by Ptolemy (see, for example, Eliade 1985, 63f.). Patricia Crone (1987, 134–137) challenged this consensus by noting that Macoraba lay in the wrong place and its etymologies were implausible, while none of the Late Antique sources at hand referred to Mecca nor to the Quraysh (see also Morris 2018, 35f. and 42f.). Crone’s scepticism not only severed the link between Macoraba and Mecca, but cast general doubt on the assumption that Mecca was a major node on the incense route (Crone 1987 and 2005; for a critique, see Bukharin 2010). In other words, not only did Macoraba lie in the wrong place (for being Mecca), but so did Mecca (for being the birthplace of Islam). Since Crone’s intervention, the economic, philological, and architectural arguments put forward against Mecca as the birthplace of Islam concern the lacking importance or virtual absence of pre- Islamic Mecca on the Arabian Peninsula in pre-Islamic and non-Islamic sources before the mid-eighth century. Revisionist authors also draw on the Qurʾānic Arabic “dialect,” which would better fit Syria, or they argue with the miḥrābs of early mosques, which point to different directions and only began pointing to Mecca in the eighth century (e.g. Crone 1987; Kerr 2014a; Gibson 2011). Thus, the emergence of Islam is said to 174 have most probably occurred not in a “barren place” (Crone 1987, 6f.) like Mecca, but farther to the North, beyond the peninsula, in Petra, Gaza, Syria, Mesopotamia, or even in Merw in the eastern part of the Persian Empire (Hawting 1999; Gibson 2011; Crone 1987; Nevo /Koren 2003; Kerr 2014a, Ohlig 2009). What unites revisionist scholars is their doubt regarding the Ḥijāzī origins of early Islam. They suggest that traditions that refer back to early Islamic Mecca were possibly fabricated in ʿAbbāsid times to give an Ḥijāzī orientation to events that probably took place outside it; and they conclude, from indirect evidence, that “the notion of an early Meccan framework cannot be attested before the first half of the second [Muslim] century” (Ibn Warraq 2007, 225, quoting Bashear 1989, 232). Such a fabrication thesis holds that the centre of the origins of Islam was relocated (from an unknown place) to Mecca after a period of 150 years.

A further argument which follows from these revisionist approaches is that the Arab-Islamic expansion from the Arabian Peninsula to Andalusia, Central Asia, and India, from the first third of the seventh up to the mid-eighth century, did not happen by military force but mainly peacefully. The indirect argument is that most of the adventurous and heroic deeds depicted in the Arabic accounts of the genre of futūḥāt (“conquests”) literature were more or less fabricated and lack archaeological evidence (Nevo/Koren 2003; Donner 2010, 106–144; Bangert 2016, 86–89, 530–535, 746).

In this respect, a further argument holds that the Oriental Christian dissenters of the Council of Chalcedon (451 CE), the Nestorians and Monophysites, welcomed Muslim rule because all of them agreed on the human nature of Jesus and thus preserved an anti-Trinitarian and Jewish form of Christianity against an emerging Greek dogmatism. The Council of Chalcedon had fixed the Trinitarian dogma and ruled that the divine nature of Jesus Christ was united with His fully human nature “unconfusedly, unchangeably, indivisibly, inseparably” (Bindley 1899, 226). The Christian disunity in the Byzantine Empire helped the Persian and Zoroastrian Sassanids, who tolerated Oriental Christians, to conquer Jerusalem, Damascus, Egypt, and parts of Anatolia and Armenia at the beginning of the seventh century. Therefore, this narrative suggests that the emperor of Constantinople might have given up on Oriental Christians and ceded the Near East to his former (Christian) Arab allies. A further suggestion in this context is that the hijra calendar, whose beginning is dated to 622, has nothing to do with Muḥammad’s emigration from Mecca to Medina, but with the beginning of Emperor Heraclius’ counter-offensive against the Sassanid Empire. To summarize, most of what the Muslim sources recount about the early history of Islam seems to be a forgery from ʿAbbāsid times in this view⁸ (for an overview see Bangert 2016). 175

Revisionist scholarship does not build up a comprehensive alternative theory for the emergence of Islam; it rather experiments with different approaches that question established chronologies, locations, and narratives of early Islam in contradistinction to traditional Muslim and non-Muslim scholarship. The mistrust of early Arabic sources leads to a deconstructive reading of early Arab accounts and later Orientalist reconstructions, which, for example, leads to post-dating Muḥammad’s death and thus changing the chronology of the events in early Islam (Shoemaker 2012). Such attempts to reorganize the early history of Islam according to an alternative reading of sources are enmeshed in vague evidence. It is a kind of a paradox that revisionists cast doubt on the writings of Muslims from the second and third centuries AH⁹ because they were as far from Muḥammad’s as we are from Napoleon’s times today, while they try to convince us that their own speculations, 1400 years after the events, are scientifically and linguistically more trustworthy.

As there are hardly any sources available for the first 150 years AH, it is certainly true that we actually cannot know much about the historical Muḥammad, although “we probably know more about Mohammed than we do about Jesus (let alone Moses or the Buddha)” (Crone 2008). Since the historical reliability of Muslim sources, mostly written after 750 CE, is doubtful, it is nearly impossible to extract historical kernels of truth in reports about the life of Muḥammad. In spite of questioning the value of these sources altogether, an alternative scholarly approach therefore tries to determine which traditions about a certain event are the earliest, with the intention to reconstruct the genesis of the corpus of Islamic tradition (e.g. Motzki 2000). Yet as radical sceptics neither accept nor practice this approach, the approach and its results have become part of a scholarly debate about methodological questions (Shoemaker 2011; Görke et al. 2012; Görke/Motzki 2014; Schoeler 2014).

Critique of Revisionist and Mainstream Arguments

In spite of its bold assertions, (neo-)revisionist scholarship has not solved the fundamental uncertainties in studies about the emergence of Islam and is, in this respect, surprisingly similar to conventional scholarship, which it so severely criticizes. Revisionist scholarship is based on at least three problematic assumptions that it has not overcome to this day. Firstly, casting doubt on the traditional chronology of the history of early Islam—post-dating or pre-dating certain events and constructing 176 a different historical context—is certainly a useful scholarly challenge; however, revisionist scholarship has not given birth to a convincing alternative chronology for early Islamic events. Secondly, insofar as revisionist scholarship indulges in a radical re-interpretation of Qurʾānic texts and early Arabic inscriptions by re-constructing relations to non-Arabic sources and terms, it remains a philological exercise producing thought-provoking insights. Yet as we know next to nothing about the social, religious, and ritual practices during the first century after Muḥammad’s presumed death in 632—if we put aside the Arabic texts produced another century later—, then these speculations rest on thin ice. Especially meagre is evidence for the idea that early Muslims understood themselves as Jewish or anti-Trinitarian Christians or a combination of both and read liturgical texts of Christian origin that were only later adjusted to the Islamic Qurʾān; or that the people who adjusted and punctuated the text misread or misunderstood most of it, whether intentionally or unintentionally. Thirdly, re-reading the Qurʾān with a more comprehensive (not exclusively Arabic) philological toolbox is obviously worthwhile, but it has so far only led to the creative re-construction of Christian textual layers beneath the Qurʾānic text—by changing the punctuation of Arabic letters or creating new meanings of Qurʾānic terms with the help of non-Arabic vocabulary. However, an “Urkoran” (Lüling 1993), the hymnal or liturgical text (possibly) in Syro-Aramaic on which the textus receptus of the Qurʾān (or parts of it) is presumably based has not been detected. In contrast, the statement still holds that “the Qurʾan is not borrowing or retelling the Biblical story, but rather commenting on it” (Reynolds 2010b, 585). Griffith (2013a; 2013b), who has tried to determine when the Bible became an Arabic scripture, has shown that “the Bible is at the same time everywhere and nowhere in the Arabic Qurʾān; there are but one or two instances of actual quotation” (Griffith 2013a, 2). While the number of what counts as actual quotations from the Bible is certainly debatable, there is a consensus that Biblical material must have circulated orally in Arabic at first; written translations into Arabic were made by Christians and Jews outside of Arabia, and only after the Arab conquests (Griffith 2013, 3).¹⁰ This suggests that the codification of the Qurʾān initiated or pushed the translation of the Arabic Bible.

The whole debate about the “dark origins” of Islam with its two opposing camps and their sometimes polemical tone—represented by the so-called Saarbrücken school and the Inârah Institute around Karl-Heinz Ohlig, Gerd-R. Puin, and Luxenberg, on 177 the one hand, and the Corpus Coranicum group in Berlin around Angelika Neuwirth, on the other hand—cannot be dealt with in further detail here. For the purpose of our discussion on the emplacement of (early) Islam, it is, however, important to note that authors from the opposing camps create different spatial contexts in their text-centric approaches for a more or less context-less Qurʾān and construct causality via chronology. This creates several problems visible in (1) the concept of Late Antiquity; (2) the container understanding of space; (3) the quest for historical kernels of truth in literature or space; (4) the attribution of genealogical and exegetical meaning to Mecca and Medina in Qurʾānic studies; (5) the geographies of Jewish and Christian influences; (6) the question of the in/stability of Islam resulting from internal divisions and its rapid spread; and (7) the search for the Christian roots of anti-Trinitarianism.

(1) While a consensus seems to exist that it is important to locate the emergence of Islam in the context of “Late Antiquity” (Brown 1971 and 1978; Fowden 2015 and 2016; Schmidt et al. 2016a; Shoemaker 2014), this historiographical term is employed to mark different spaces and elaborate different explanations. The historian Garth Fowden (2015; 2016) uses the first millennium as a periodization that allows drawing broad strokes of the history of Islam—emergence, development, maturity—and integrating it into a general history of Euro-Asia by shifting the focus away from the Mediterranean to the East. In contrast, research on the origins of Islam and on the Qurʾān (e.g. Neuwirth 2010a; Neuwirth et al. 2010) has a much narrower understanding of Late Antiquity. Here, the term is delineated as a primarily virtual “thinking space” (Schmidt et al. 2016a), which helps to grasp the transformation processes in Late Antique thought as well as the transfer of biblical, post-biblical, philosophical, and Jewish knowledge to the Arab space (Schmidt et al. 2016b, 21). Situating early Islam in a transcultural context called Late Antiquity thus helps to fill the blank of the first Islamic century with different—mainstream and revisionist—agendas. While both camps try to explain the Qurʾānic entanglement with biblical material—the “Biblical subtext” (Reynolds 2010a) of the Qurʾān—, they envision the impact of Judaism and Christianity differently (see also (4)). In a paradoxical turn, the Late Antique contextualization of Qurʾānic studies prioritizes a focus on similarities, entanglement, and parallels with Judaism and Christianity; yet both mainstream and revisionist scholarship proposes that the Qurʾān either actively reworked or altered the Jewish and Christian impact. In consequence, mainstream scholars tend to see early Islam as an already stable entity, thus underlining the non-epigonic singularity of Islam’s unique treatment of biblical motives and downplaying inter-religious parallels. Revisionist scholars cherish the idea that early Islam was a rather unstable and weak entity, yet strong enough to alter, if not falsify, biblical motives. Beyond this, neither the temporal nor the spatial boundaries of the Late Antique contextualization of the Qurʾān are clear. Interestingly 178 enough, Neuwirth (2010a) announces her study, entitled the “Qurʾān in Late Antiquity,” as a “European approach” in the subtitle. This example is disturbing insofar as it blurs the boundary Islam/Europe in the title, but re-erects it in the subtitle. It aims to move the history of Islam closer to late antique Europe, yet distances the academic study of the Qurʾān from a non-European perspective. Not making these differences visible is questionable; spatializing them, however, suggests that the history of Islam in Late Antiquity is located in a “thinking space” transgressing boundaries, whereas its study requires taking sides.

(2) The revisionist approach questions the importance of Mecca/Medina as the places of revelations and takes them for literary fiction and a later construction (e.g. Kerr 2014a; 2014b). Mainstream scholarship clings to the traditional differentiation, going back to Theodor Nöldeke (d. 1930), that distinguishes three phases of Meccan revelations from Medinan Sūras and attributes meaning to this emplacement of the Sūras (see Nöldeke 1970 and (4)). In both cases, a container understanding of space is dominant, according to which early Islam spread from smaller to bigger containers. In the first case, Islam moves from Mecca to Medina, subduing the Arabian Peninsula, then Mesopotamia, and so forth (e.g. Berger 2016). In the revisionist case, a deconstructivist mode radically questions the starting-point and simply plays with the idea of putting early Muslims in another location, from where they spread. From the technical side, this is reminiscent of Kamal Salibi’s reverse attempt (1985; for a critique: Beeston 1988), in which he used place names of the Hebrew Bible and epigraphic evidence to locate ancient Jewish history in the Ḥijāz and ʿAsīr, arguing that there was a severe mismatch between Jewish stories and archaeological findings in Palestine, while the names made perfect sense in Arabia.

(3) Mainstream and revisionist scholars share the desire to distinguish kernels of historical truth from fabricated traditions. The paradox in this respect is that revisionist scholarship generally opposes mainstream approaches that try to separate the “real” kernels from the fictious parts of early Islamic literature; yet it proceeds in the selfsame way by trying to identify the original place of the origins of Islam and distinguish it from the fabricated one (i.e. “Mecca”).

(4) The mainstream approach, which departs from a refined differentiation between Meccan and Medinan Sūras (e.g. Neuwirth 2010a), ascribes exegetical meaning to space. This differentiation is based on the different styles of the Sūras, although Meccan and Medinan verses have also been inserted to otherwise named Sūras, and the Fātiḥa (Qurʾān 1) is even understood as both a Meccan and Medinan Sūra (Bobzin 2010, 603n10). Moreover, Neuwirth’s (2010a, 2010b) main view is that the composition of the Qurʾān reveals a dialogical character. However, the assumption that the Qurʾān expresses a “dialogue” between Muḥammad and his community and that 179 its oral composition preceded the written one is based on a textual interpretation that presupposes a causality as well as social practice in Mecca and Medina for which there seems to be little empirical evidence (for a polemical critique, see Groß 2014, 803–931).

For example, Neuwirth interprets the Qurʾānic references to the story of Abraham as different “lectures” in Mecca and “re-lectures” in Medina and thus tries to corroborate a chronology of the Qurʾānic text as well as a chronology of an inner-Qurʾānic re-interpretation (Neuwirth 2016, 191–196; also Sinai 2009). The relevant verses stem from the Sūras 2, 3, 14, 21, 37, 52, 53, 57, and 87. Neuwirth (2016) holds that these verses actually reflect dialogues that really took place “in” Mecca or Medina. She holds that the re-lecture of Abraham’s story in Medina finally managed to localize Abraham’s sacrifice in the Kaʿba and in Mecca. Neuwirth’s view creates problems on the spatial and textual level. In the first step, Neuwirth (2016) does not differentiate between the physical-historical space in which a dialogue takes place and the symbolic ascription of spatial markers to Sūras; spatial markers that are otherwise missing in the Qurʾānic text were later introduced for the Sūras and mainly follow from a stylistic analysis or from the literature of “occasions of revelation” (asbāb al-nuzūl) of later centuries. In a second step, the text-immanent ascription of space to different parts of the text is used to establish a historical chronology of the textual genesis. In a third step, the re-lecture in physical Medina expresses an exegetical development; this development is reflected in the location of Abraham’s story in pre-Islamic Mecca, which is said to have caused the establishment of new ḥajj rites in Mecca when the Prophet was in Medina, although the Qurʾānic text does not mention either Mecca or the Kaʿba explicitly in this context, while ḥajj rites and their alteration by Muḥammad are also mentioned in the traditional literature for the pre-Medinan time.

The circular reasoning in three steps amalgamates the genesis and exegesis of the text as well as the physical and symbolic dimensions of space. Although Qurʾānic verses are characterized by different styles, and their revelation is spatialized in the Islamic tradition as well as in the Orientalist approach following Nöldecke, it might still be possible and useful to differentiate between the symbolic ascription of space to verses and the physical location of their revelation or, in other words, between the styles of the written Qurʾānic text and the “dialogues” of believers. Although a chronological ordering of the verses which make up the story of Abraham along the axis of Meccan and Medinan revelations is not implausible, certain requirements result: the style of the revealed verses must correlate with their chronology; the chronology of the revealed verses must constitute a causal connection; the written verses must give a complete picture of the oral lectures and dialogues so that the “re-lecture” in Medina adds something to the “dialogues” that have already taken place in Mecca; the community must have been aware of the links between dialogues held at different times; believers 180 must have understood these links and adopted a new rite; in spite of these links, the relevant verses were later allocated in scattered places in the textual corpus of the Qurʾān; this scattering of chronologically and causally linked verses in the Qurʾān was possible because the believers either knew or did not know the links; the exegetical literature re-affirmed the previously existing links; and today’s scholarly reconstruction of these links does not create an illusionary chronology or causality.

(5) The different geographies of the opposing camps often induce an emphasis on different “influences”: one side implicitly underlines Jewish influences, while the other often stresses Christian ones. The strategy to downplay Christian influences by highlighting Jewish ones—thus opposing the revisionist thesis of a Syro-Aramaic lexis (Luxenberg 2004) and a vernacular poesy (Lüling 1993) in the Qurʾān—seems problematic. When Sinai/Neuwirth (2010, 19) argue that the Psalms are “the only biblical corpus that has exerted a formative impact” on the theological and literary shape of the Qurʾān, this assumption entails a formal and a spatial problem.

On the formal level, the focus on Islamic-Jewish similarity downplays the homiletic quality of the Qurʾān. Reynolds (2010a, 243–258), who is not convinced by Neuwirth’s parallelization of the Qurʾān and Pslams, speaks out in favour of sorting the Qurʾān into the larger homiletic tradition because of its insistence of not delivering a new message (Qurʾān 3:3, 5:48), its permanent allusions to biblical material, and its topical wandering. The apparent random character and the seemingly arbitrary organization of the Qurʾān can thus be explained as a structural trait and a reflection of the Syriac homiletic tradition. Reynolds (2010a, 254) argues that our understanding would gain much from reading the Qurʾān’s unique treatment of biblical motives as well as the Qurʾān’s biblical subtext in the light of a homiletic tradition, and that “the Qurʾān itself points us to this reading.”

The spatial problem lies in the fact that our knowledge of the history of Jewish and Christian practices and doctrines in the Ḥijāz is rather limited. Mazuz (2014), for example, has tried to discover the religiously and spiritually “authentic” life of Jews in Medina by indirect evidence drawn from early Arabic sources; he concludes that most Jews were “Talmudic-Rabbinic Jews in almost every respect” (ibid., 99). This remains speculative because the “customs, laws, practices, and beliefs of the Medinan Jews are essentially constructed in absentia” (Hughes 2015, 581). Thus, the question arises of why “it was so important to Mazuz that the Jews of Medina be normative” (ibid., 582). On the one hand, Mazuz creates “a continual and ‘authentic’ Jewish identity that stretches out from the ashes of the destruction of the First Temple and moves directly through to the codifiers of the Babylonian Talmud (and beyond)” (ibid.). On the other hand, Mazuz supports the suggestion that “since Muhammad did not have access to the Talmud, the Jews of Medina function metonymically for Talmudic law” (ibid.). In this 181 sense, the Muslim community’s “dialogue” (Neuwirth) could also involve Talmudic Jews and thus their texts, no matter whether this dialogue found expression in Meccan or Medinan Sūras.

Characterizing Meccan parts of the Qurʾān as “hymns” that “sound like distant echoes of the Psalms” (Neuwirth 2010b, 735) might, however, not only overstate similarity and neglect differences (Sanni 2015, 310); it also envisions the echo “in” Mecca before a possible dialogue with Mazuz’s “Talmudic Jews” “in” Medina could have taken place. In this respect, it seems that Neuwirth lumps together the spatial and temporal differences in the term of a “distant echo”, whereas in other instances, a painstaking differentiation between physical Mecca and Medina extracts symbolical meaning from a probable sequence of events at these places (see (4) above). Would it go beyond the scope of a geographical location of the revelation in Mecca/Medina if the echo of the Psalms stemmed from Yemen or from beyond the Arabian Peninsula? Another possibility of explaining the source of a Jewish echo in Mecca is to fall back on the speculations of the outsiders Dozy (1864) and Lüling (1985; 2000). Dozy (1864) argued that the Simeonite tribe of Israel had defeated a part of the Canaanites in an unidentified town called Ḥormah, which Dozy related to the sanctuary at pre-Islamic Mecca. According to him, the Simeonites had conquered the land in which Mecca would emerge, and they were called Ishmaelites in the sense of “immigrants” (see Morris 2018, 25). Lüling (1985; 2000) further elaborated Dozy’s thesis by arguing that the cult in pre-Islamic Mecca was controlled by Levite emigrants after the Jewish conquest (Morris 2018, 27). Both authors’ narratives try to explain parallels between Muslim and Israelite rituals “as the residues of an Israelite conquest” (ibid., 25).

(6) The different revisionist and mainstream emplacements of early Islam show that the entanglement of Judaism and Christianity with early Islam—and the contradictory negotiation of these relations in the Qurʾān—still forms a fundamental challenge in the studies on the origins of Islam. The multi-religious landscape of Arabia is mainly spelled out as having varied impact on or commonalities with Islam, rather than as a panorama of doing and undoing differences. Since entanglement has been a recurrent topic in religious polemics as well as in scholarship throughout the centuries, mainstream as well as revisionist scholarship actually navigates known waters by stressing the close relationship of the three religious traditions, yet re-writes their entanglement in different ways. Mainstream scholarship—especially in Neuwirth’s case, discussed above—tends to view Islam as a stable entity from early on, while revisionist scholars place Islam more clearly in terms of dependency on Christianity and Judaism, which finds its expression also in moving the location of Islamic inception to distant (or nearer?) places. The revisionist idea that “Islam is little more than the sum of earlier monotheistic parts” (Hughes 2017, 871) is part of a lasting discussion on the origins of 182 Islam by non-Muslims and Orientalists. The problem with this idea is not only that it suggests that the Qurʾān and Islam are epigonic, but that it assumes “that Judaism and Christianity were somehow more stable than what was gradually coalescing into Islam” (ibid.). Neuwirth’s insistence that the Qurʾān is not epigonic (in spite of its diverse borrowings) and that it possesses a full-fledged theological concept (which requires complex re-constructions of the text) treats Islam as an entity as stable as the other two traditions—in spite of multiple internal divisions, which were a subject of conflict in the first century AH, according to Islamic tradition.

Even Donner’s (2010) proposal, which sees early Islam as an ecumenical movement and tries to bridge the gap between revisionist and mainstream scholarship, follows the same paradoxical trajectory. It holds that early Islam began as a movement of “believers” (muʾminūn), including “pious Christians and Jews” (ibid., 171), and gradually transformed into a group of clearly distinct muslimūn. It is not only possible that this proposal describes “a tolerant world that may not have existed” (Crone 2010). It also nurtures the suspicion that it blurs the boundaries between (Islamic) discourse and reality (Zelletin 2016, 128). Generally, Donner’s study is based on two important distinctions. He understands early Islam as a religious movement rather than an economic, social, Arab, or political one—as most of revisionist scholarship might prefer to do—, and he envisions how rather stable forms of Judaism and Christianity “gave birth to an unstable Islam,” as Hughes (2017, 882) put it. The first problem with this proposal is that we do not know much about “either the contours or contents of Christianity or Judaism, in large part because we have very little idea of what any of these three ‘religions’ would have looked like at that particular time and in that particular location” (Hughes 2017, 882). The second problem is that unstable Islam—riven by two civil wars and rivalling groups in its first century—was strong enough to rapidly conquer vast parts of Christian Byzantium and all of Zoroastrian Persia because of their military exhaustion and lacking social cohesion due to internal religious diversity and strife. This paradox remains a main challange in the historiography of early Islam (see also Berger 2016).

(7) With regard to pre-Islamic Arab Christians in and around the Arabian Peninsula,¹¹ the generally accepted wisdom today is that they might have been heretical groups after the Council of Chalcedon (451 CE), which built and maintained churches and monasteries at several places on the peninsula (Finster 2010, 70–75). Revisionist scholars tend to further assume that Muslims evolved from Judeo-Christians who opposed the Hellenization and rationalization of the Christian doctrine that became apparent with the dogmatic formulation of God’s Trinity, which took shape between 183 the First Council of Nicaea (325) and the Eleventh Council of Toledo (675). Against this background, there has been, among Western scholars, “a never-ending hunt for a particular Christian sect” (Reynolds 2009, 252) that could explain Islamic anti- Trinitarianism, which—given the variety of Christian groups—often led to puzzlement.

Although the Qurʾān refuses the Trinity in several verses (Q 4:171, 5:73, 6:163, 17:111, 25:2, 37:35, 47:19, and 112:1–4) (see Bobzin 2015, 58–66) and often understands it as an expression of polytheism, scholars have often pointed at the Qurʾānic account of the crucifixion of Jesus, which is only mentioned in two verses (Q 4:157–158).¹² Most of the Islamic tradition and most Western scholars of Islam take it for granted that the Qurʾān denies Jesus’ death, assuming “that this denial reflects the influence of Christian docetism” (Reynolds 2009, 238). New approaches, however, suggest that the verses can be understood, in line with the prevalent Christian perspective, as denying neither his death nor his elevation by God (Reynolds 2009; Lawson 2009). By studying forty Qurʾān commentaries from the seventh to thirteenth centuries as well as modern exegetes, Lawson (2009) has tried to show Qurʾānic “neutrality” in this respect, arguing that it was tafsīr, not the Qurʾān, which denies the crucifixion of Jesus with different arguments. Most interpreters tried to make sense of shubbiha lahum (“it appeared to them”), a hapax legomenon in verse 4:158, by reading it as “what was seen crucified on the cross was just an image: a phantom, not the real Jesus or perhaps even a substitute” (Lawson 2009, 3). Reynolds (2009, 240) argues that the relevant Qurʾānic verses can be read in the sense of “God (and not the Jews!) first made Jesus die, and then made him ascend to heaven.” Therefore, Reynolds (2009, 255) thinks that the Qurʾānic passage makes two points: First, it is about Jewish infidelity (since Jews killed Jesus just as they did other prophets) and, second, about God’s control over life and death. In this sense, the Jews who claimed to have caused what appeared to them as the death of Jesus “are twice in error” (ibid.). Instead of establishing a main difference between Islam and Christianity, the passage rather seems to be in line with the mainstream of Christian anti-Jewish rhetoric (ibid., 255–258).

The Qurʾān’s “neutrality”—as opposed to the bias in tafsīr literature—might remain a question of debate (Shah 2010, 198), since the Islamic exegetical tradition also created the historical context for the Qurʾān, as Reynolds (2009, 252) admits.¹³ Given 184 that this particular Qurʾānic passage does not directly express an anti-Trinitarian credo but was only interpreted as such, the claim that the Qurʾānic wording might reflect the credo of a particular Christian sect remains dubious, although Christian Monophysitism might have influenced its interpretation. Yet from a Jewish point of view, a non- Trinitarian understanding of God is not at all peculiar. Thus, the Qurʾānic passage about the Jewish treatment of Jesus can be understood as a way of establishing a difference to Jewish, not Christian, views. In any case, scholarly attempts to localize the origins of anti-Trinitarianism in a Christian sect indirectly address the question of Islam’s originality. Behind the disproportionate attention to Islam’s anti-Trinitarianism lurks no new insight, but the old idea of Islam as a Christian heresy, which reflects “tropes and stereotypes that have dominated Western Christian discourse about Muslims since the reception of John of Damascus’ depiction of Islam as a heresy” (Ralston 2017, 756) (see further below).

A Topological Approach

In most scholarly approaches about early Islam, (geographical) space merely forms the more or less plausible background scenery of events and contacts; it does not establish a relational space, which comes into being through human action, imagination, and emotion. The absence of Mecca in pre-Islamic documents and maps is no proof for the fictionality of pre-Islamic and early Islamic Mecca or the improbability of the emergence of Paleo-Islam¹⁴ in Mecca; it is rather a hint at the fact that the emergent Islamic community produced Mecca and its growing importance in a historical process. From the point of view of spatiality, the obligation to explain a complex phenomenon such as the emergence of Islam exclusively by its rootedness in a singular location or at the crossroads of civilizations is void. The revisionist attempt to re-date and re-locate the origins of Islam is based on a conception of Mecca that is at the same time too narrow and too wide—similar to the inverse cherishing of the conventional idea that there was a Meccan and Medinan period of the Qurʾānic revelation. It is too narrow if it restricts the meaning of Mecca to its geographical boundaries, and too wide if it marks the emergence of such a variegated phenomenon that was later called Islam. On the one hand, Mecca is bigger than itself because it has historical, political, and symbolic extensions; like with all geographical sites, the question is what the term “Mecca” means and where it ends. On the other, the emergence of Islam is a multi-layered process that transcends narrow geographical boundaries as well as a Meccan 185 period, however defined. If we understand origins as a rhizome (rather than roots) and locations as spaces characterized through movement, connection, and circulation (rather than as immobile places), then the topology of early Islam must be relegated neither to a barren place nor located at the crest of cross-fertilization. Instead, the space of early Islam can be understood as polycentric, dynamically changing, uneven, and hierarchically structured. Depending on whether the focus lies on political, economic, religious, historical, imaginary, or emotional dimensions, the term “Muslim space” induces different outlines, movements, perceptions, representations, and spatial practices.

The relative absence or presence of Mecca and Medina in different parts of Islamic culture and practice is due to different contexts and their history—with regard to the Qurʾān, sīra and tafsīr literature, cartography, qibla, and ḥajj. The following remarks are meant to underscore the obvious point that the “emergence” of Islam does not go back to “dark origins” and was not completed by the first century AH, neither in a political or military nor in a doctrinal or cultural sense. Debating the origins of Islam only makes sense with the spatial and temporal extensions of Mecca and Medina in mind. Although we do not know much about early Islamic Mecca and Medina, what is certain is that these places changed decisively as their commemorative, representative, architectural, and physical landscapes were reworked in the course of time, maybe prior to and during the emergence of Islam, but certainly after it. The production of sanctity for these cities took winding and contradictory roads and was connected with a wide range of activities involving literary production, investment, organization, urban planning, and even destruction.

Following these considerations, I highlight nine aspects that are connected with the spatiality of early Islam: (1) the relative absence of space in the Qurʾānic message; (2) the emplacement of Muḥammad’s biography in the Western Arabian Peninsula; (3) the composition of the Qurʾān with extensions beyond the Arabian Peninsula; (4) the processes that created a highly symbolic and representational value of Mecca, Medina, and the Kaʿba; (5) the establishment of ḥaram districts; (6) the connection between Islam’s spatial expansion and Muslims’ reorientation in time and space; (7) the position of Mecca in a web of places and routes; (8) the conquests, spread, and re-invention of Islam beyond the first century; and (9) the emergence of Islam as an object of global knowledge production.

(1) The nearly total absence of names and places in the Qurʾān can be seen as a fundamental theological concept. Whereas names and places attribute evidential value and conclusiveness to stories and narrations, the Qurʾān renounces this concept of affirming truth claims. It is rather concerned with the question of how one can discern that its message is God’s revelation. In its self-referential style (Wild 2006), it 186 comments on prophets and messengers but does not tell a straight story either about Muḥammad’s activities or about the course of the world since its creation.

(2) Both these aspects are dealt with in later biographical and historiographic works, in which Muḥammad and Mecca/Medina are placed within the course of world history and treated as exceptional. Although these reports claim to depict historical events, the places mentioned in this kind of literature do not necessarily reflect geographical realities. As part of a salvation history, they emplace the revelation and link it with the actions of the Arab prophet as well as with prophetic time; therefore, they have high symbolic, imaginative, and emotional value. Although the geographical and symbolic dimensions of a place may be relatively independent, it is rather improbable that they are totally detached from each other, as revisionist scholarship assumes. A topological approach understands the production of space (Mecca) and its sacredness (Kaʿba) as two intertwined historical processes that mutually confirmed each other, but in relation to competing places and claims (see more below under (4)). Thus, on the one hand, the weight of the Meccan framework, which has imposed itself since the second century AH, can simply be explained by the fact that the production of a Meccan centrality took some time (no matter where the exact (!) place of origins lies); an alternative explanation seems superfluous. On the other hand, not only do time and place work against each other in the resettlement thesis, but the historical, mythical, and symbolic dimensions of Mecca would have been torn apart at a certain stage and would only have merged at a later stage, without ever wholly mending the break. It seems highly improbable that (a) the different processes that produced the historical and symbolic dimensions of a place named Mecca were, at one point in time, suddenly interrupted, and that (b) we can again disentangle them today. Even given that archaeological excavations hinted at another historical “Mecca,” such hints (c) would still have difficulties explaining how the location or relocation of the Kaʿba and the sīra worked and (d) would be overshadowed by the symbolic weight of today’s Mecca, that would stand in contrast to archaeological evidence from a forgotten place. The resettlement thesis—in the absence of excavations—therefore requires a higher amount of credo quia absurdum than the production of a pilgrimage site in a barren place.

(3) The question of historical authenticity is not only relevant for the sīra, but also for the reports about the collection and codification of the Qurʾān (see e.g. Motzki 2001; Schoeler 2010). However, it seems clear that the geographic places of these events, which led to the creation of the Qurʾān as a book, is different from the way in which the Qurʾānic revelation is emplaced in sīra literature. Even according to the Islamic tradition, the collection and codification of Qurʾānic material took place under circumstances that point beyond the Ḥijāz. Thus, the collection was not restricted to Medina and Mecca under the caliphate of ʿUthmān, but famously included other 187 collections by Ibn Masʿūd, Ubayy, and Abū Mūsā in Kufa, Damascus, and Basra, among others (e.g. Reynolds 2010a, 208).

(4) Although the importance of Mecca and Medina stems from the fact that they represent central places in the traditional literature about the emergence of Islam, they are involved in early events in different ways and to different extents. Therefore, the important symbolic as well as representative role that Mecca and Medina came to play has been produced through a complicated historical process (Peters 1986; Munt 2014). By all accounts, Mecca and Medina have never been regarded unanimously as the single most important places of early Islam, nor the hub of the world. In the religious imaginary, they were part of a web of interconnected places and spaces. From a political point of view, the centre of Arab power moved from Medina to Damascus, Kufa, Ḥarrān, and Baghdad in the first centuries. Mecca was no central place of political power, although the Meccan Quraysh dominated the fate of the early Muslim community; Ibn Zubayr’s reign as anti-caliph in Mecca (r. 683–692) was but a short episode. Islamic geographers of the ninth and tenth centuries did not place Mecca but Baghdad in the central position; showing Mecca as the node of the world was a phenomenon beginning only in the late tenth century CE (Webb 2013, 9).

By then, the matter of the sacred hierarchy between Mecca, Medina, and Jerusalem was “far from closed” (Munt 2014, 189), although a Muslim majority unmistakably considered both Mecca and Medina holy places by the end of the Umayyad period (Peters 1994a, 107–154). However, some scholars rearranged the ranking of the three cities, while others put forward claims of sanctity for Damascus, Kufa, Baghdad, and other cities. According to one tradition, Muḥammad’s ḥaram is located at Medina because he founded the first prayer room in his house there, while the building of the Kaʿba in Mecca is associated with Abraham (Webb 2013, 7f.). The establishment of a sacred space in Medina was, however, controversially debated among scholars, since the boundaries of the ḥaram district were presumably altered or expanded after Muḥammad’s death; other scholars even questioned the very existence of a holy district in Medina and opposed its establishment (Behrens 2007, 211–226). At the same time, caliphs encouraged pilgrimage (ziyāra) to some sites in Medina although there was already an early scholarly debate about the permissibility of performing prayers at the Prophet’s grave (ibid., 227–276). Although the later writings of scholars and historians filled Medina’s sacred topography with places, it is questionable whether the landscape they wished to depict really matched the physical conditions (Munt 2014, 96). For anti- Shīʿī Sunni scholars, who regularly frowned upon the local rulers in the Ḥijāz because of their Shiite leanings and connections (Ende 1997), the cosmological and symbolic meaning of Mecca and Medina was certainly always clearly distinguishable from the existing topography and the pro-Shīʿī inhabitants and tendencies of these places. 188 The emergence of the Kaʿba and the sacred district in Mecca is even more complex. Although the Kaʿba seems to have existed already in pre-Islamic times, its founders, its establishment, and the cult practices are a riddle, and different theories connect it to paganism or Jewish and Christian forerunners (Peters 1986, 104–122; Hawting 1982; Busse 1993; Dozy 1864; Lüling 2000). From the ninth century CE onwards, Muslim historians introduced the idea that the original building of the Kaʿba went back not only to Abraham but to Adam, who also performed the first ḥajj (Webb 2013, 8–9), and even that the Kaʿba existed prior to God’s creation (Antrim 2012, 43–48). Accordingly, after the expulsion from Paradise, Adam came down to earth in Mecca, where he met Eve, who had come over from India. The spatial significance of Mecca, which “lies on a blessed trajectory directly beneath the throne of God” (Webb 2013, 9), explains why God “sent down”¹⁵ Adam here and why pilgrims must circumambulate this place in parallel to the angels’ praise of God.

Early Muslim accounts are rather outspoken about structural changes and a pre- Islamic “pagan” use of the Kaʿba (Rubin 1986), which allowed Muḥammad to “restore” the original rites. Accordingly, the Quraysh rebuilt the sanctuary with the help of Muḥammad before his prophethood and thereby re-discovered the well of Zamzam, a treasure, and swords (Wheeler 2006, 19–46). In a report, a Coptic craftsman is said to have directed the re-construction of the Kaʿba (Peters 1986, 111). The Quraysh are also said to have removed the Black Stone from its original place on a nearby hill and put it into the Kaʿba four years before Muḥammad’s first revelation (Rubin 1986, 120). The other sacred stone, maqām Ibrāhīm, showing an imprint of Abraham’s foot according to Islamic tradition, was transferred, together with the Black Stone, and placed adjacent to the Kaʿba or inside it at various times, while it is placed inside the mosque today (ibid., 122f.). When Ibn Zubayr, the anti-caliph, declared his sovereignty, the Umayyads layed siege to Mecca and the Kaʿba was destroyed by fire. Ibn Zubayr, who claimed to have found Abraham’s original founding stone, had the whole building destroyed and rebuilt with the intention to restore its Abrahamic condition. When the Umayyads retook the city, they destroyed the new structure and returned the Kaʿba to the form it had had when Muḥammad and the Quraysh worshipped there (Peters 1986, 113). While the Kaʿba seems to have remained unaltered after that, the surrounding ḥaram district became the object of enlargement and monumentalization. The extension of the ḥaram and properties around it became “a prime object for a new aristocracy of Muslims” (Peters 1986, 114).

There is evidence that “the Kaʿba was not always the primary place of worship in Mecca” in pre-Islamic times (Rubin 1986, 118), and even less a central destination 189 for pilgrims from the whole Peninsula. Thus, it seems that the rise of the Kaʿba and Mecca as the paramount sacred site and pilgrimage destination in Islamic times was “a gradual phenomenon” (Munt 2014, 189; c.f. Peters 1994a, 107–154). Although there is only sparse evidence of social and ritual practices in early Islamic centuries, we can assume that these practices changed over time. With regard to the ḥajj, the Umayyad and ʿAbbasid caliphs tried to use it for the legitimization of their power and monopolized leading the ḥajj. Yet we also know that political opponents and some Sufis scorned the ḥajj rites. Most famously, the Qarmatians went as far as killing pilgrims at the Kaʿba, throwing their dead bodies in the well of Zamzam, and robbing the black stone in 930 CE. Pilgrimage, however, was certainly not yet a mass phenomenon. Later reports often mention the participation of religious scholars, princesses, and princes, while the appearance of rulers was seen as an exception (Möhring 1994). Even when the organisation of pilgrimage caravans became an elaborate and expansive operation, Mecca remained a small or modest-sized town before its massive expansion in the twentieth century.¹⁶ While traveller Johann Ludwig Burckhardt (d. 1817) estimated the number of Mecca’s inhabitants at 30,000 and pilgrims at 90,000 in 1814, today’s numbers have risen to 1.5 million inhabitants and approximately 3 million pilgrims. Due to heavy investment in infrastructure, and in accordance with a desacralizing view of the Islamic heritage, Saudi authorities have not only massively restructured Mecca and Medina, but also destroyed 300 historical sites in both towns during the last two decades alone, an estimated 95 percent of the ancient Meccan heritage (Botz-Bornstein 2015, 165).

The history of the cemetery of Baqīʿ in Medina clearly expresses different readings and arrangements of the sacred, of Islam, and of space (Bobeck 2018). As many ahl al-bayt and ṣaḥāba (relatives and companions of the Prophet) as well as four Imams, caliph ʿUthmān, and respected scholars of Islam are buried in Baqīʿ, it has been a central place for Sunnī and Shiʿī visitors for centuries, respected and enlarged by Umayyad, ʿAbbasid, Selcuk, Mamluk and Ottoman rulers. The Saudi-Wahhabi conquests, however, led to acts of desecration in 1806 and 1925/26 in which shrines and gravestones were destroyed. After the Ottoman re-conquest of the Ḥijāz in the nineteenth century, the shrines could be re-erected only temporarily. As Ende (1997, 318) has put it, “protests against the destructions at Baqīʿ in general and of the tombs of the Imams in particular have been leitmotiv of Shiite writings about Medina” since 1925. For the sacred geography of the Shīʿa, whose emergence Haider (2011) described already for the Kūfa/Najaf of 190 the second century AH, Baqīʿ is not only an integral part of the collective memory, but also a central place in the web of interconnected shrines that leads believers to friendly sites and away from inimical zones. From a Wahhābī understanding, visiting a tomb conflicts with tawḥīd (the unity of God), thus verging on idolatry and shirk (polytheism). To avoid the temptation of idolatry, Saudi-Wahhabi spatial practices aim at disrupting the relation between the living and the dead (Bobeck 2018, 6). Not only the practices in Baqīʿ are strictly controlled today; Wahhabi scholars still discuss the destruction of the so-called Green Dome (al-qubba al-khaḍrāʾ), erected by the Mamluks in 1279 CE over the graves of Muḥammad, ʿAlī, Abū Bakr, and ʿUmar. Thus, Saudi practices aim at the desacralization of (formerly) sacred places in Mecca and Medina. Yet they also support a kind of “sacralization” of other places (ibid., 14, 17–19): the birthplace of Muḥammad b. ʿAbd al-Wahhāb (d. 1798), the founder of the Wahhābiyya, in Dirʿiyya has been turned into a tourist attraction and museum (ibid., 14).

(5) The establishment of ḥaram districts in the Ḥijāz must have had a direct effect on Jews and Christians living on the Arabian Peninsula because they were forbidden to enter or to reside in Mecca and Medina. The prohibition is based on a saying by Muḥammad¹⁷ as well as on accounts of early Muslim scholars stating that in the time of the second caliph, ʿUmar b. al-Khaṭṭāb (d. 644), the Ḥijāz or the whole Arabian Peninsula was freed from the presence of non-Muslims through expulsion (Munt 2015, 250).

However, Muslim scholars debated the concrete implementation of this ruling for several centuries, and the evidence of expulsions is questionable. The existence of Jewish and Christian communities within the Ḥijāz and on the Peninsula was reported by various Muslim sources throughout the early Islamic centuries; in Yemen, the Jewish presence lasted up to the twentieth century (ibid., 251, 259–261). Jurists’ rulings in the early centuries also varied considerably regarding which places non-Muslims were allowed to enter (ibid., 257f.). As scholars assumed that all non-Muslims had been expelled, they constructed or altered the geographical definition of Ḥijāz/Arabia, in the sense that Ḥijāz/Arabia could be only where non-Muslims were not living (ibid., 263). A source-critical study has trouble identifying the historical background of reports on Muslim persecutions of Jews in Medina at the times of the prophet Muḥammad (Schöller 1998). The picture given by reports depicting the massacre of up to 900 men of the Jewish Banū Qurayẓa in Medina is far from clear; it is difficult to establish whether the expulsion of Jews from Medina really happened or to what kind of historical event these reports might refer. However, the importance of this trope in the cultural memory is attested by the fact that two of the three illustrations about the Medinan period in the Khalili manuscript of the Jāmiʿ al-tawārīkh (1306–1311 CE), which was composed by the 191 Jewish convert Rashīd al-Dīn al-Ḥamdānī (d. 1318), show Muhammad’s raids against the Banū Nāḍir and the Banū Qaynūqaʿ, respectively (Hillenbrand 2014). In any case, the partial and gradual implementation of the ḥaram districts is another hint at the fact that the emergence of the sacred Islamic centre seems to have been a gradual process (Munt 2015, 251). The prohibition for Jews and Christians to stay in the Ḥijāz, and the difficulty to implement this prohibition, implicitly affirm the right of Jews and Christians to reside in the, however defined, rest of the lands under Muslim control.

(6) When non-Arabic and Arabic sources from the mid-seventh century CE onwards talk about the origins of Islam (see, for example, Hoyland 1997), we have to take into consideration the simple fact that Arabs and Muslims already controlled a vast territory, extending from Europe to India and Central Asia at that time, while Arab traders had reached ports in China, although they might not yet have built mosques there.¹⁸ If Islamic empire building had not been successful, there would have been no demand for written accounts about the origins of Islam, be it in Arabic or non-Arabic languages. Muslim historians of the eighth and ninth centuries CE were not only concerned with adjusting biblical stories to their own narratives, but they already drew on the history of China, India, and ancient Egypt as a pre-history (Donner 1998, 127–134). Later examples in the genre of universal history include chronicles written by Christian authors who depicted history from Adam to their lifetime, for example the historian Jirjis b. al-ʿAmīd (known as al-Makīn, d. 1273 CE) from Egypt, in his Majmūʿ al-mubārak (1262–1268) (e.g. Frenkel 2015, 83–91), or the Syrian-Orthodox Bishop Gregory Bar Hebraeus (d. 1286) from Persia, who wrote in Syriac and Arabic (e.g. Todt 1988). Another famous example is the chronicle Jāmiʿ al-tawārīkh (1306–1311 CE), by the Jewish convert Rashīd al-Dīn, which is considered the first written world history (Melville 2008) not only because of the breath of topics, from Adam to the Mongols, but also because its author combined a variety of sources to depict the histories of the Chinese, Arabs, Persians, Indians, Jews, and Franks.

From this consideration follows a challenge for historiography and spatial imagination, because the search for the origins of Islam is always already a back projection in which earlier and later events as well as distant regions and Arab centres are interconnected. With regard to the historiographic challenge, even early accounts about the emergence of Islam are tinged with an understanding of the later events of the Arab-Muslim expansion that had already changed Muslims and Islam, both of which had emerged under conditions that were, in all probability, different from those in 192 which the accounts about their origins and expansion were written. With regard to the spatial challenge, the dissolution of boundaries and the crossing of frontiers went hand in hand with processes of empire building—the strengthening of a centre in Baghdad and regional sub-centres— and with processes of legitimatization, in which accounts of Arab conquests obviously played an important role. These processes also induced a geographical and historiographic re-orientation to the past and to the Arabian Peninsula. Phenomena like the introduction of the miḥrāb in mosque architecture and the growing importance of the ḥajj (see the next point) can probably be understood as a physical expression of an interlocking between the Arab expansion and the Muslim re-connection to the past and the Arab Peninsula.

(7) In mainstream and revisionist views of early Islam, the importance accredited to Mecca and Medina stems from the question of how or whether Islam spread from there to other places. From a topological point of view, the importance of these places rather stems from the routes taken to them and from their (changing) relations to other places. Mecca and Medina not only attracted Muslim travellers like Nāṣir-i Khusraw (d. 1072/78) from Central Asian Merw (ḥajj in 1050), Ibn Jubayr (d. 1217) from Valencia (ḥajj in 1184), and Ibn Baṭṭūta (d. 1368/77) from Morocco (ḥajj in 1326) (Peters 1994a, 71). Since the early sixteenth century, the cities of Mecca and Medina also stirred the curiosity of European travellers and writers, who had to dress and behave like Muslims and profess Islam when they visited them (Peters 1994b, 206–265). The Italian Ludovico de Varthema (d. 1517) saw Mecca and Medina in 1503, and portrayed them in detail as the first European in 1510; Joseph Pitts (d. 1735?) entered Mecca presumably as a slave and depicted the ḥajj rites in his report of 1704. In the nineteenth century, Johann L. Burckhardt, Richard F. Burton (d. 1890), and Charles M. Doughty (d. 1926) published their popular travelogues in 1829, 1857, and 1888, respectively.

Moreover, Peters (1994b, 71) highlights that there also exists a body of geographical writing in Arabic that is committed to describing the social, historical, and architectural dimensions of the Holy Lands of Islam. He particularly emphasises its roads and stations: “The very earliest example of the genre, by Ibn Khurdadhbih (d. 893–894), is in fact called The Book of Routes and Provinces, as are many of its successors, including those of Istakhri (ca. 951) and Ibn Hawkal (ca. 977)” (ibid., 71f.). Thus, since the ninth century, Muslim narratives about places underline their connectivity, linking the centrality and universality of such places as Mecca, Medina, Jerusalem, and Baghdad with their boundedness and particularity (Antrim 2012, 33–60). Petersen (2018, 87– 121), who explores the Chinese-language Islamic texts collected in the Han Kitab of the eighteenth century, has shown how the meaning and practice of the ḥajj changed from a Chinese-Muslim point of view over the centuries. By comparing the works of the three Sino-Muslim authors Wang Daiyu (d. 1658), Liu Zhi (d. 1724), and Ma Dexin (d. 1874), 193 he elucidates that the perception of the ḥajj turned from “a symbol of true belief, to a potential critical practice, and finally to an essential observance and religious duty” (ibid., 87).

That Mecca and Medina were not the centres of political power for most of their history also helped to turn them into a cosmopolitan meeting point for scholars, pilgrims, and refugees. Although the local rulers of the Ḥijāz had to come to terms with different overlords throughout the centuries, the strategically peripheral location spared them attacks since the times of the Qarmatians up to the beginning of the nineteenth century. Standing outside the formal hierarchy of the Ottoman Empire, Medina was in a position to become one of the most important centres of Muslim scholarship from the sixteenth century onwards, as Reichmuth (1998; 2000) has shown. There, a Kurdish theologian could come into contact with scholars from Timbuktu, have a pupil from Sumatra, and write a commentary to a work by an Indian colleague. Or an Indian scholar, studying in the Ḥijāz, could build up a scholarly network reaching from Western Africa over the Maghreb to India and Central Asia. The growing presence of scholars from Daghestan and the Western Sahara since the seventeenth century illustrates the flowering and consolidation of Islamic culture and Arabic scholarship in these regions as well as a growing awareness and appreciation among Arab Muslims for these groups and regions.

A turning point in Mecca’s and Medina’s history was the moment when European shipping companies started to offer ḥajj travels by steamships in the nineteenth century and “applied the same business logic and mechanisms that they deployed in the transportation of other populations” (Miller 2006, 192; also Peters 1994b, 266–300; Slight 2017). The organisation of the ḥajj to Mecca began changing massively, and the numbers of pilgrims from South Asia jumped to unknown heights at the beginning of the twentieth century (Miller 2006). When air travel was introduced, and procedures for pilgrims from all over the world further standardized, in the mid-twentieth century, the lines between pilgrimage and tourism were further blurred and the effect of globalisation transformed a pillar of Islam into a religious, yet highly commercial, form of tourism (McLoughlin 2018).

(8) The spread of Islam and the movements of Islamic conquests did not end with the Arab expansion of the first century AH. Rather, a series of conquests and re-conquests by various tribal, ethnic, and religious groups from different places occurred in the following centuries. These include the movement of the Arab Banū Hilāl across Northern Africa to Mauretania from the tenth century onwards; the re-conquests of the Maghreb and Muslim Iberia by Berber groups, the Almoravids and the rival Almohads, in the eleventh and twelfth centuries; the conquests of the originally non-Muslim Mongols, who helped spread Islam further across Eurasia; and the conquests of the Ottomans, 194 who took the chance to subdue big swaths of already “Islamic” land in the fifteenth and sixteenth centuries (Reinkowski 2016). Especially the Mongolian impact, whose origins are potentially as “dark” as the Islamic ones, is worth mentioning, because the originally non-Muslim Mongols helped spread Islam to new regions in Central Asia, India, Russia, and Poland, intensified the trans-regional traffic of goods, travellers, and refugees, and produced a common space of interaction, imagination, and destruction (Jackson 2017).

In this broader sense, it is certainly one-sided to attach the emergence of Islam to a singular place like Mecca/Medina when there were multiple, polycentric, and continuous efforts of new beginnings in Islam. There are many examples, ranging from different Sūfī orders to the Wahhābiyya, that show how movements of religious renewal spread from different places. A case in point is the Naqshbandiyya-Mujaddidiyya, founded by Aḥmad Sirhindī (d. 1624) in India (Reichmuth 1998; 2000). As Sirhindī and his adherents broke with the previous mystical universalism in India, relations to the Moghul emperors became strained, so that initial support turned into prosecution. While Sirhindī’s major opus was officially forbidden in 1679, and a fatwā from the Hijāz posthumously declared him an unbeliever in 1682, the order spread in the Ottoman Empire and was officially recognized in the aftermath of the disastrous campaign against Vienna in 1683, which ended the predominance of the Kadizadelis at the Porte. Beginning in the eighteenth century, the Mujaddidiyya further spread to Central Asia, Russia, Egypt, and North Africa. Finally, it helped to bring about the political and administrative reforms in the Ottoman Empire of the nineteenth century and played a role in the coterminous disempowerment of the Janissaries and the Bektāshīs, a rival Sūfī order.

(9) The importance attested to revisionist scholarship on the previous pages does not stem from its alternative readings of Islamic origins. It rather serves as another example of the temporal and spatial extension of what is termed as “Islamic origins,” since the distant past is connected to present research and definitions of Islam originate in a global arena, thus attributing the understanding of pre-modern places in Arabia to modern places such as Caen, Saarbrücken, or Berlin.

From a topographical and topological point of view, the revisionist re-location of the birth of Islam closer to the heartlands of Christianity and Judaism seems to be an unnecessary operation. It is well established by now that the Arabian Peninsula was connected with the surrounding world in pre-Islamic times and that Arabs themselves had already spread beyond the peninsula (e.g. Hoyland 2001). The northern and southern parts of Arabia fell, at different times, under Roman, Byzantine, Persian, and Ethiopian rule. The trade routes running from South Arabia in pre-Islamic and Islamic times connected distant places with each other. The incense route ran from Dhofar 195 by land or by sea to Jeddah, Gaza, Damascus, and Petra,¹⁹ while sea trade connected India to Egypt. Arabian culture outside the Peninsula can be traced back to the Bronze Age (Hoyland 2001), and even paleogenetics suggests that Arabs were migrating to the North of the Peninsula and returning to it long before Islam (Tadmouri et al. 2014, 396f.).

The geographical rapprochement of early Muslims to Christian and Jewish lands in order to explain their entanglement follows neither from a spatial argument nor from philological detail work alone. It reflects tropes that have dominated Western religious and scholarly discourses about Islam for centuries.²⁰ Lüling (1993, ix–xii, 15–23; 2007) is quite explicit in this respect because he not only dedicated his work to the reformed theologian Martin Werner (d. 1964), but also subscribed to the latter’s view that the doctrine of the Trinity was not part of Early Christianity but a Hellenistic, Roman- Imperial “distortion” of Jesus’ self-image (Lüling 1993, ix). Lüling’s motivation was to prove that the Early Christian non-Trinitarian understanding of Jesus was “exactly” reproduced in the Qurʾān. He held to have “irrefutably” proven that the Qurʾān and the history of the emergence of Islam was falsified in the first two centuries by the emerging Muslim “orthodoxy,” which also “fundamentally” re-interpreted the historical figure of Muḥammad (ibid.). Against this background, he asked Christianity to concede that Islam alone preserved the correct Early-Christian understanding of Jesus, while he demanded that Islam accept that its own genesis was forged in a profound sense (ibid., ix–x). He understood his research results as prolegomena for ecumenical dialogue between the religions, because the dogma of the Trinity posed the main point of difference between Christianity and Islam (Lüling 2007, 297). As he aligned himself with the rationalism of the Enlightenment, the historical-critical school, and a Christian theology critical of church dogmata, he also understood his research results as the “fulfilment” of the critical Qurʾānic studies of the last two centuries (ibid., 301f.). He had few positive words reserved for scholars of Islamic studies since the second half of the twentieth century, because in his view, they had resumed neither the critical research of liberal Protestant theology nor the liberal Qurʾānic studies of the nineteenth and early twentieth centuries. Lüling’s personal background was that he began to study theology and specialised in the Hebrew Bible, but could not hope to become pastor because of his anti-Trinitarian stance. Therefore, he transferred to the department of Islamic studies, where his qualification work of 1969 (extended version in 1973/74 and 1993) met with determined resistance—in spite of some non-public encouragement— 196 so that he lost his academic position and had to end his university career, while his research was not discussed in Germany for several years (Lüling 2007, 300–306); it only found belated recognition and reception by Anglo-Saxan scholars after he had published his findings in English (Lüling 2003).

In this respect, the questions and doubts about the emergence, redaction, and composition of the Qurʾān have less to do with the nature of Islam itself than with the question of the purpose of critical research and academic debate (for a critical Muslim view, see Bagrac 2010). Rivaling viewpoints on the origins of Islam not only express a lasting circulation of known tropes and their rearrangement, but also serve as a means to negotiate one’s own scholarly self-understanding about the purpose and aim of academic research.

Conclusion

Where do the multi-religious origins of Islam lie and what do they tell us? Trying to answer this question, I have firstly identified the historiography of the emergence of Islam as a “wicked problem” that resists a clear and simple resolution. I have tried to pinpoint tricky and malicious problems in revisionist and mainstream scholarship on the previous pages. The focal point was the controversial spatiality of early Islam and its entanglement with Judaism and Christianity. My intervention, although not aiming at a consensus, intended to bring Islamic tradition, mainstream scholarship, and revisionism into discussion with each other and turn divisive certainties into perplexities. In this respect, I have tried to show that the attempts of mainstream and revisionist scholarship to defend or relocate the birthplace of Islam follow from a limited understanding of space. Classical Orientalist scholarship tried to distinguish the “true” historical kernel on the origins of Islam in early Arabic literature—an attempt that revisionist scholarship rejects with regard to Arabic sources, yet it follows the selfsame approach with regard to the primordial places of Islam. Thus, revisionism is marred by a self-contradiction, while mainstream scholarship in Qurʾānic studies seems to be trapped in a confusion between the physical and symbolic dimension of space and between the genesis and exegesis of Qurʾān, especially with regard to Mecca and Medina.

After this review of the recent literature, I have secondly argued that we should understand the birthplaces of Islam neither as a literary “invention” nor as a given fact. I have argued for their social and historical construction, thus opposing total arbitrariness as well fixed immobility in location, shape, and meaning. My point of view is that Islam did not come into being “in” Mecca or any other spatial container, but that 197 Muslim actions and mobility produced the space of early Islam physically, symbolically, and emotionally. I have further argued that Islam—as a full-fledged religious worldview— emerged neither in a barren place nor at the crossroads of civilizations. Instead, its formation happened through a longer, polycentric historical process. By following a topological approach, I have tried to elucidate that the constitution of places called Mecca and Medina and the attribution of historical, symbolic, imaginative, emotional, and representative quality to them were intertwined processes. I have further suggested that the production of the sacrality for Mecca and Medina in turn ascertained the general multi-religious topology of Islam.

Thirdly, I have concluded that the Islamization, sacralisation, and growing importance of Mecca and Medina were gradual historical processes in a multi-religious landscape—although we have only limited information in the sources about spatial and social practices in early Islam. However, these processes of sacralisation involved the restructuring of the Kaʿba, the Ḥaram districts, and the adjacent quarters as well as investments in infrastructure and security, mirroring the growing importance of ḥajj and ziyāra. It also seems to have been beneficial for the emergence and preservation of Mecca’s and Medina’s sacred character that both cities were remote from the political centres of Islam for most of their history, which allowed them to thrive as centres of religious scholarship and retreat in the shadow of political turmoil. From here, the Islamic and sacred character of Mecca is understood as a product of spatial practices that have evolved over time and distributed meaning partly in contradictory ways. For example, the Islamic character of archaeological sites in Mecca did not save them from destruction by an Islamic government; and the growing number of pilgrims to Mecca and Medina moved the ḥajj closer to tourism.

I have taken issue with the consensual use—in an otherwise differing scholarship— of a somewhat ossified concept of Late Antiquity that originally was an opener to integrate Islam into a larger vision of global history (e.g. Brown 1971; 1978; Fowden 2015; 2016). In studies on the Qurʾān and early Islam, this concept either proceeds from entanglement to reach uniqueness or envisions cross-fertilization, yet argues with the early in/stability of religious entities whose contours, practices, and norms are mainly unknown to us. My counterargument was to focus on internal diversity, dynamic development, and re-negotiation. The in/stability of what is called early Islam must be understood as a relational phenomenon in comparison with the in/stability of similar umbrella terms, such as paganism, Judaism, and Christianity in the same period (see Al-Azmeh 2014). Early Arabic-Muslim sources give a complex and dynamic picture of rivalries and alliances between different tribes and groups in which differing religious views did not always play the most important role. 198

Drawing on this argument, my fourth point is that religious diversity in Arabia tells us exactly this—that diversity lay at the cradle of Islam, existed before its birth, and lingered on after it. This multi-religious topology of Arabia is directly connected with the problem of defining what “religion” (or “Islam”) is. Taking into account Smith’s (1982, xi) argument that “religion” emerges through second-order acts of classification and therefore is “solely the creation of the scholar’s study,”²¹ means to take the problem of classification seriously. When a scholar describes Islam as an anti-Trinitarian form of Christianity, then the scholar’s boundary work implicitly creates a division as well as a common denominator. At the same time, this operation marks and unmarks what Islam is and establishes closeness as well as distance to Judaism and Christianity. Mainstream as well as revisionist scholarship on the birth of Islam often connects chronology with causality and treats Islam as the result of various religious impacts. In contrast, a topological approach is critical of the lasting “influence of influences” (Tauber 2018) and rather looks at the human practices that constitute X and Y as distinct categories to establish a chronology as well as an authoritative relation between them. Therefore, it is important to note that (1) “the Bible is at the same time everywhere and nowhere in the Arabic Qurʾān” (Griffith 2013a, 2); (2) Jewish and Christian traces on the Arab Peninsula are attested in literature and archaeology, although we do not know exactly what kind of Jews and Christians were there; and (3) Arabs had far-reaching contacts beyond the Arab Peninsula long before the advent of Islam. These aspects suggest the existence of physical and imaginative landscapes with a diverse and multi-religious character. As the earliest sources show, the actors seem to have negotiated this diversity by various practices and discourses. Thus, early Islam was not only entangled with other religious traditions and cults, but also linked to opposing practices and discourses—ranging from the acceptance of religious plurality (as expressed in the so-called constitution of Medina) to fields of ambiguity (as expressed in debates about the presence and the rights of Jews and Christians in the Ḥijāz) to tendencies of purification (as expressed in narrations about the expulsion of Jews from Medina or inner-Islamic conflicts). As these practices and discourses of “un/doing difference” (Hirschauer 2014) lingered on in later centuries, it seems questionable to ascribe either “dark origins” or a full-fledged theological concept to Islam of the first century AH when we have a thriving culture of Muslim debate that explores, time and again, the origins and the meaning of Islam up to this day.

Therefore, regarding the geographical rapprochement of Paleo-Muslims to Christian and Jewish lands in revisionist scholarship, my argument was that it is supported neither 199 by a spatial point of view nor by philological detail work alone. It reflects tropes—Islam as an anti-Trinitarian heresy— that have dominated Western religious and scholarly discourses about Islam for centuries. This brings us back to question of the sitz im leben of Qurʾānic studies and studies on early Islam; they are located inside the hierarchies of knowledge production that often serve political projects to the disadvantage of the colonial or post-colonial “other”. My contribution questions the way in which the heterogeneous research on the origins of Islam has attained growing importance against the background of current popular debates on Islam. The popular negative image of Islam certainly poses an additional, new challenge to the scholarship of early Islam, since populist critics of Islam also tend to eclectically draw on critical research.

Against this backdrop, it seems legitimate to ask what drives the new ambitions of research on the origins of Islam, as we seem to make only little actual progress and probably know more about the historical Muḥammed a priori than about Jesus, Moses, or the Buddha, as even Crone (2008) has admitted. Maybe the trope from a Greek text, written between 632 and 634, that “a false prophet has appeared among the Saracens” (ibid.) is still key to understanding the subliminal forces that guide at least parts of the debate. The trope of falseness seems to make a difference. Whereas the riddle of whether Buddha’s supreme wisdom came into this world in the sixth, fifth, or fourth century BCE is mainly discussed by academic scholars,²² the fierce debate about the beginnings of Islam suggests that some participants believe it necessary to defend Islam while others think that Muḥammad’s forgery is still waiting to be fully unmasked.

Andrew Rippin (2012) recently pondered about the negative Muslim reception of “secular, academic scholarship” (ibid., 2), which he also called “Euro-American scholarship,” although he immediately recognized that this kind of terminology “is fraught with difficulty” (ibid., 5). He then takes himself to task because he falls into the trap of the underlying cultural assumptions of the West versus Islam: “It is a fact of the modern world that religious allegiance does not correspond to geographical location in any sense” (ibid.). Yet the appeal of such a faulty kind of boundary work stems from the fact that what is seen as a (Western) “polemic put forth in the guise of academic research” (ibid., 3) often triggers a (Muslim) “apologetic mode of response,” which is framed as a critique of the modern/postmodern world “that does not ascribe an ultimate value to belief in the divine” (ibid.). Decidedly Muslim approaches then try to reconcile research with belief in the Qurʾān as a book “in which there is no doubt” 200 (referring to Qurʾān 2:2) and often end up in “sweeping arguments which can neither be proved nor disproved and must be seen as part of broader protest against Western domination” (ibid. 4, quoting Waardenburgh 1993). As Rippin admits, the “West versus Islam” classification is questionable, since Muslim as well as non-Muslim scholars are involved in Islamic and Qurʾānic studies in a critical and constructive manner; the challenge for everyone, he writes, lies in “putting negative images behind us and dissociating ourselves from work that is not worthy of being called scholarship” (ibid. 6).

However, the nature of serious scholarship is a question in the scholarly debate itself, as I have tried to show on the previous pages. The controversial debate about the nature and origins of Islam is a case in point. It serves rivalling scholars as a way to substantiate their respective understanding of the meaning and aim of historical-critical research and works to stabilize scholarly hierarchies as well as camp thinking. In this sense, the debate so far has often had the tendency to be a spatial practice and boundary work in its own right. It should be opened up for new transgressions.

Literature