A direction of prayer is a falsifiable thing. A mosque is a building; its\nqibla[1]\n wall has an azimuth; that azimuth can be measured with a\ncompass and a satellite fix and compared against the great-circle bearing to\nany point on Earth. Unlike a doctrine, a wall cannot be reinterpreted into\nagreement. It either points where the tradition says it should, or it does not.
\nOn the largest survey yet made of the oldest datable mosques — the work of the\nCanadian researcher Dan Gibson, assembled over two decades and published as\nEarly Islamic Qiblas (2017) and Let the Stones Speak (2023) — a great many\nof them do not point where the tradition says they should. For roughly the\nfirst Islamic century, their qibla walls do not face the Hijazi city now called\nMecca. They face north, toward \nPetra\n,\nthe abandoned Nabataean capital in the sandstone gorges of southern Jordan.
\nThat is one claim, and it is contested. What makes it worth a long essay is\nthat it does not stand alone. It arrives braided together with three others:\nthat Mecca is missing from the ancient geography of Arabia; that the earliest\nrecorded measurements of the Ka'ba describe an irregular four-sided structure\nrather than a cube; and that the language and the very alphabet of the Qur'an\ndescend from the Aramaic of the Nabataeans rather than from any dialect of the\nHijaz. Underneath all four sits a fifth observation, the one the Wheel of\nHeaven corpus cares about most: that the earliest sources do not call the new\nreligion Islām at all. They call it the Ḥanīfiyyah — the religion of\nAbraham.
\nThis Explainer works through the strands one at a time, because each is\nindependently contestable and the argument lives or dies on whether they\ngenuinely converge. It then hands the floor to the hypothesis's most serious\nopponent, the historian of Islamic astronomy David A. King, whose alternative\nexplanation has to be taken seriously before any of this is believed. Only then\ndoes it read the whole through the \nWheel of\nHeaven\n frame — not as a finished conclusion, but as a hypothesis that,\nif it holds, sharpens a pattern the corpus already follows: the recurring\nrecovery of a lost Abrahamic original.
\nBegin with the gap the whole subject lives inside. The events at the origin of\nIslam are dated to the early seventh century CE. The earliest surviving Islamic\nliterary sources that narrate those events — Ibn Isḥāq's biography, the legal\nand historical compilations — were written down, in the forms we have, roughly\ntwo centuries later. Between the founding generation and the first detailed\naccount of it lies a long stretch in which the written record is thin to the\npoint of silence.
\nGibson's organising move is to treat that silence as an opportunity for a\ndifferent kind of evidence. During those two centuries the early community was\nnot writing the histories we would like to have, but it was building. Mosques\nwent up across an empire that ran from Spain to Central Asia, and a mosque\nrecords, in its foundations, a decision that no later editor went back to\nrevise: which way to face. Where the manuscripts fall silent, the masonry keeps\ntalking. The orientations of the early mosques are a documentary record laid\ndown at the time, by the builders themselves, immune to the editorial layering\nthat reshaped the literary tradition.
\nThe qibla, in other words, is a control on the texts. If the buildings and the\nbooks agree, so much the better for both. If they disagree, the buildings have\nthe stronger claim to record what the early community actually did, because they\nwere not rewritten.
\nGibson's survey classifies the early mosques by where their qibla walls aim.\nThe dataset — published openly as the online Qibla Tool, so the orientations can\nbe checked against satellite imagery — sorts them into a handful of groups:\nmosques that face Petra; mosques that face the Hijazi Mecca; mosques that face a\npoint between the two; and mosques oriented parallel to a line drawn from\nPetra to Mecca. He sets a tolerance band of ten degrees around each target and\nclassifies anything outside every band as \"unknown,\" rather than forcing it to\nfit a story.
\nThe pattern he reports is chronological. The earliest mosques, from the first\ndecades of Islam, face Petra. The Mecca orientation appears later. In between\ncomes a cluster he calls the \"between\" qibla, which he dates to around 87–88\nAH[2]\n and associates with the governor al-Ḥajjāj ibn\nYūsuf[3]\n — a transitional, confused phase in which mosques\nsplit the difference, as though builders had been told the holy city had moved\nand were uncertain where it now lay. By Gibson's reckoning the Petra qibla\npersists from the outset of Islam until roughly 132 AH, overlapping for decades\nwith the emerging Mecca orientation before the southern city wins out.
\nTwo examples carry the texture of the claim. In Oman, the early Al-Midhmar\nmosque at Sumaʾil is oriented within about two degrees of Petra — a precision\nthat, over such a distance, is hard to dismiss as coincidence. And in Jerusalem,\nthe al-Aqsa mosque preserves the shift fossilised in place: its qibla wall, on\nGibson's reading, was built facing Petra, while the prayer-rugs and the\nworshippers inside are angled off the wall's axis to face Mecca, so that the\ncongregation prays at a slight diagonal to the building that houses them. The\nstructure and the practice point in two different directions, because the\nstructure is older than the change of mind.
\nThe single most important external check on this came from outside Gibson's own\nwork. The statisticians Walter R. Schumm and Zvi Goldstein took his raw\norientation data and ran them through formal analysis, asking whether the early\nqiblas behaved like a set of measurements aimed at a target or like noise. Their\npeer-reviewed conclusion was that, under Gibson's assumptions, the early mosque\nqiblas are statistically consistent with intentional orientation toward Petra,\nwith the scatter widening as distance from the target grows — exactly the\nsignature one expects from a real but imperfect aiming method applied over\nincreasing range. The statistics do not prove the history; they test whether the\ndata have the shape the history requires, and they find that they do.
\nA reasonable reader stops here with an objection, and it is the same objection\nKing presses hardest: a builder in Córdoba or Samarkand cannot see Petra. He has\nno spherical trigonometry, no satellite fix, no way to compute a great-circle\nbearing to a point a thousand miles beyond the horizon. If the early Muslims\ncould not have aimed accurately over such distances, then the apparent precision\nof the Petra qiblas must be an illusion, and the whole edifice falls.
\nGibson's answer draws on the navigational and surveying methods that were\navailable in late antiquity, recovered partly from the work of King himself on\nearly Islamic timekeeping. The core technique is the Indian Circle, a\ngnomon-and-shadow method that requires no mathematics beyond drawing arcs. A\nvertical rod is set in a level patch of ground; the tip of its shadow is marked\nin the morning and again in the afternoon where the shadow touches a circle\nscribed around the rod's base; the line joining the two marks runs true\neast–west, and its perpendicular runs true north–south, to within a fraction of\na degree. From an accurate set of cardinal directions, a builder who knows the\ndirection of a distant place — preserved as tradition, as a bearing carried by\ntravellers along the well-trodden caravan roads, or as a star that rises over it\n— can lay out a wall to that bearing with surprising fidelity. The method\ndegrades predictably with distance, which is exactly why Gibson's far-flung\nmosques in Morocco and Central Asia scatter more widely around Petra than the\nnearby ones do, and why Schumm and Goldstein found error growing with range\nrather than staying flat or random.
\nThe point is not that the early builders were geometers. It is that orienting a\nwall toward a remembered distant place is a solved problem of practical surveying\nthat long predates Islam, and that the error-with-distance pattern in the data is\nthe fingerprint of such a method honestly applied. King reads the same inability\nto compute great circles as proof that the builders could not have aimed at\nanything precise and so fell back on his many star-and-wind schemes. The dispute\nis not over whether seventh-century people could do spherical trigonometry — both\nsides agree they could not — but over whether a non-mathematical aiming tradition\ncould carry a single bearing across an empire. Gibson says it could, and points\nto the Indian Circle; King says it could not, and points to the proliferation of\nfolk schemes. The data sit between them.
\nIf the holy city was Petra, then the obvious objection is the obvious city: what\nabout Mecca? Was it not an ancient sanctuary in its own right? Here the argument\nturns from masonry to the documentary geography of antiquity, and the claim\nsharpens into something startling — that there is no secure evidence for Mecca's\nexistence as a city before Islam at all.
\nThe strongest pre-Islamic candidate has always been a single line in Ptolemy's\nGeography[6]\n: a place in Arabia called Macoraba. For\ngenerations the equation ran almost on the strength of the consonants —\nMacoraba begins with M, contains an r and a b, therefore Macoraba is\nMecca, therefore Mecca is on a second-century map. Gibson and, independently,\nthe historian Ian D. Morris dismantle the inference from both ends. Morris's\ndetailed survey of the scholarship concludes flatly that the identification was\nnever more than a guess hardened by repetition, and that the philological link\nbetween Macoraba and Makka does not hold. Gibson adds the cartographic\nargument: when Ptolemy's known distortions of Arabia are corrected by\nre-projecting his coordinates against rivers and sites that can be located,\nMacoraba does not fall where Mecca is, and no Ptolemaic city falls at Mecca's\ncoordinates. The city that the tradition makes the immemorial navel of the world\nleaves no trace in the ancient geography that records its supposed neighbours.
\nThis is a negative argument, and negative arguments are weaker than positive\nones — absence of evidence is not, by itself, evidence of absence. But it does\nreal work in the convergence. The qibla data ask the reader to move the holy\ncity north; the geography removes the principal reason to think it was ever in\nthe south.
\nA third strand returns to stone, and to the sanctuary itself. The word Kaʿba\nmeans \"cube,\" and the building in Mecca is, near enough, cubic. But the earliest\ndetailed measurements of the Ka'ba — recorded by the Meccan historian al-Azraqī\nin the ninth century, describing the structure as rebuilt during Muhammad's\nown lifetime — give four sides of unequal length. The Ka'ba al-Azraqī\ndescribes is an irregular quadrilateral.
\nGibson takes those four irregular measurements as a fingerprint and goes looking\nfor a structure that matches them. He reports finding one: an altar standing\nslightly off-centre before the Qasr al-Bint in Petra, a structure whose\nasymmetry had puzzled observers precisely because it does not sit where a\nsymmetrical plan would put it. On his measurement the Petra altar matches\nal-Azraqī's four irregular sides, steps included; and if the steps are removed,\nthe remaining dimensions match the present Ka'ba in Mecca. He is careful — more\ncareful here than his critics sometimes allow — to call this a candidate\nidentification rather than a proven one. The Petra structure has not been\nexcavated or studied with this question in mind, and a match of measurements is\na lead, not a verdict. Set beside the Nabataean habit of venerating deities in\ncubic standing stones, or baetyls[5]\n, it becomes a\nsuggestive lead. It does not become a demonstration.
\nThe qibla data describe a transition: Petra first, the southern Mecca later. A\nhypothesis that asks the reader to believe a holy city moved owes an account\nof how and why, and this is where Gibson's reconstruction is at once most\nambitious and most fragile. It is offered here as the connective tissue of the\nargument, plainly more conjectural than the qibla survey it tries to explain.
\nThe hinge is the second civil war (683–692 CE) and the rival caliphate of\nʿAbd Allāh ibn al-Zubayr[11]\n, who held the holy city and\nrebuilt the Ka'ba during his reign before the Umayyad general al-Ḥajjāj besieged\nand defeated him. The literary tradition places this siege at Mecca in the\nHijaz. Gibson places it at Petra, and points to physical traces: at the\nstructure Brown University's excavators named the Great Temple, the doors and\ngaps were barricaded for defence, and stores of catapult stones — the\nammunition of a trebuchet bombardment — were recovered nearby and remain in the\nsite's storerooms. On his reconstruction the Black Stone, the sacred object at\nthe centre of the rite, was carried south during the chaos of the war, around\n65–70 AH, to keep it out of the besieging Umayyad armies; and where the Stone\nwent, the holy city's identity followed — its name, its pilgrimage stations,\nand the name of its well, Zamzam, attaching to a new foundation in the Hijaz.
\nFor a period, on this telling, there were two Meccas, and the literary tradition\nthat fixed the holy places centuries later — in geographers such as Yāqūt,\nwriting around 600 years after the events — recorded only the surviving southern\none. Gibson marshals a set of suggestive secondary traces for a general\nrelocation of sacred toponyms: early Islamic reports that the town of Ṭāʾif had\nitself \"moved\" from the north; the difficulty of fitting the topographic\ndescriptions of episodes like the boycott in the canyon of Abū Ṭālib to the\nHijazi sites; the presence of early mosques excavated outside Petra, at al-Bayḍāʾ.\nEach of these is individually light, and Gibson presents them as a cumulative\npattern rather than as proofs. A reader can accept the qibla data and still find\nthe relocation narrative underdetermined — and that is a coherent position. The\nrelocation is the part of the hypothesis most exposed to the charge of building\na story to connect dots that might not be connected.
\nThere is a darker coda that the corpus notes for its bearing on the\neditorial-suppression theme. In 930 CE the Qarmatians, a sectarian movement\nhostile to the pilgrimage as then practised, sacked the southern Mecca, killed\npilgrims, and carried off the Black Stone to their own territory, holding it for\nsome twenty-two years and refusing substantial ransom to return it. Their\nopposition was theological, not financial — they returned other loot but kept\nthe Stone. Whatever one makes of Gibson's geography, the episode is a documented\ncase of the holy object being physically removed, contested, and eventually\nrestored, which establishes that such a thing was possible and that the sources\nremember it happening.
\nThe fourth strand is the one that most directly touches the Wheel of Heaven\ncorpus's interest in language, and it is, in some ways, the most robust, because\nits foundation is mainstream.
\nThe \nNabataeans\n were an Arabic-speaking\npeople who wrote in Aramaic. Their cursive Aramaic hand, ligature by ligature\nover the late-antique centuries, became the Arabic alphabet.[9]\n\nThis is not Gibson's claim; it is the settled finding of epigraphers such as\nAhmad al-Jallad and Laïla Nehmé. The script in which the Qur'an is written is a\nNabataean inheritance. That much is simply true.
\nOn this foundation Mark Durie, drawing on al-Jallad's epigraphic work, builds a\nlinguistic argument that resolves two old puzzles at once. The first puzzle is\nthat medieval Muslim philologists, certain that the purest Arabic lived on\nBedouin tongues, combed the Hijazi dialects for the language of the Qur'an and\nnever found a match. The second is that of the thousands of pre-Islamic\ninscriptions scratched across Arabia, almost none use the Qur'an's standard\ndefinite article, al-. Durie's resolution is that the language behind the\nQur'an is Nabataean Arabic: a northern Arabic in which the al- article had\nalready become standard, written by a people so habituated to Aramaic that their\nArabic rarely surfaces in the inscriptional record under its own name. Where\nthe al- article does appear before Islam, a striking share of those\ninscriptions are in Nabataean script; and the Nabataean rasm[4]\n,\nthe consonantal skeleton, repeatedly agrees with the Qur'an. In al-Jallad's\nphrase, the Aramaic script of the Nabataeans \"casts a clear Arabic shadow.\"
\nDurie himself surfaces the tension that makes this strand matter, and he does\nnot hide it. Qur'an 14:4 has God send every messenger \"in the language of his\npeople.\" If the language of the Qur'an is Nabataean in character, and a prophet\nspeaks the tongue of his own people, then — in Durie's own words — \"it is hard\nto see how this could have been Mecca,\" whose dialect was different. Durie\nfloats and then doubts the idea that trade alone spread a Nabataean lingua\nfranca to the Hijaz. The tension is real on his own terms. It is precisely the\ntension that the Petra hypothesis dissolves, by putting the Qur'an's holy city\nwhere its language already lived. The linguistic argument and the archaeological\nargument, developed by different people for different reasons, point at the same\nmap coordinate.
\nA further, more contested layer belongs here for completeness. Christoph\nLuxenberg's Syro-Aramaic Reading of the Koran argues that several obscure\nQur'anic passages clarify when their unpointed consonantal skeleton is read\nagainst an Aramaic substrate rather than forced into Arabic — and the tradition\nthat al-Ḥajjāj added the vowel-pointing raises the possibility that the pointing\nfixed an Arabic reading over an older Aramaic one. Luxenberg's method is sharply\nrejected across much of mainstream Qur'anic studies, and the corpus flags it as\nthe most speculative thread in the bundle. It shares the Aramaic-substrate\npremise with the far better-grounded Durie–al-Jallad argument, and it is\nincluded as a possibility under examination, not as support.
\nStrip away the geography and the masonry for a moment and ask what the new\nreligion called itself. The answer, in the earliest layer of the tradition, is\nnot Islām.
\nIbn Isḥāq's biography[7]\n preserves a story set before\nMuhammad's mission. Four men of the Quraysh — Waraqa ibn Nawfal, ʿUbayd Allāh\nibn Jaḥsh, ʿUthmān ibn al-Ḥuwayrith, and Zayd ibn ʿAmr — concluded that their\npeople had corrupted the religion of their father Abraham, that \"the stone they\nwent round was of no account; it could neither hear, nor see, nor hurt, nor\nhelp,\" and resolved to go and find the true faith. In Ibn Isḥāq's words they\n\"went their several ways in the lands, seeking the\n\nḤanīfiyyah\n, the religion of Abraham.\" The\nword recurs seven times in the passage. One of the four, Waraqa, is the same\nChristian relative of Khadīja who, in the canonical narrative, later confirms\nMuhammad's first revelation. Another, Zayd, is told by a monk in the Balqāʾ that\n\"the time of a prophet who will come forth from your own country... has drawn\nnear. He will be sent with the Hanifiya, the religion of Abraham.\"
\nA ḥanīf[8]\n is one who has turned from idolatry to the\nupright, primordial monotheism; the Ḥanīfiyyah is that religion. And the\nQur'an itself uses exactly this vocabulary to define what Muhammad brings. It is\nnot presented as a new thing. It is presented as the recovery of the oldest\nthing:
\n\n\nAbraham was neither a Jew nor a Christian, but he was an upright one\n(ḥanīf), a muslim, and he was not of the idolaters. (Qur'an 3:67)
\n
\n\nSay: Rather, the religion of Abraham, the upright (millat Ibrāhīma ḥanīfan).\n(Qur'an 2:135)
\n
\n\nThen We revealed to you: Follow the religion of Abraham, the upright one.\n(Qur'an 16:123)
\n
The Qur'an claims Abraham as older than the traditions that quarrel over him,\nand assigns Muhammad's mission to Abraham's own religion. Gibson's contribution\nis to argue that Ḥanīfiyyah was not merely a theme but the movement's original\nname, with Islām — \"submission\" — coming to dominate later, as the community\nturned from converting Arabian polytheists to confronting the established\nmonotheisms. He offers the obvious parallel: the earliest followers of Jesus\nwere \"the Way\" before they were \"Christians,\" named so only later and elsewhere\n(Acts 11:26). Whether or not the name-shift was as clean as the analogy\nsuggests, the underlying point stands on the Qur'an and Ibn Isḥāq directly,\nindependent of any spade turned at Petra: the new religion understood itself as\nthe restoration of the religion of Abraham.
\nAnd the religion of Abraham, in this telling, ran through a particular son.\nIbn Isḥāq sets the four seekers' quest in the frame of the holy city, the\npilgrimage, and the altar of Abraham — the altar that Abraham was said to have\nbuilt with Hagar and their son Ishmael. The pilgrimage rites themselves are\nnarrated as re-enactments of what Abraham did: the circumambulation of his\naltar, the running between the watering-places in memory of Hagar's search for\nwater. Whatever the geography, the self-understanding is unambiguous. This was\nthe \nAbrahamic\n line of Hagar and Ishmael,\ncircling back to recover a religion it believed had been lost.
\nHere the essay crosses from reporting a revisionist hypothesis to reading it\nthrough the Wheel of Heaven corpus — and the crossing has to be marked, because\nthe two are not the same kind of claim. Gibson's archaeology is contested\nempirical work that will be settled, if ever, by excavation and by the response\nto David A. King. The corpus reading is interpretation laid over that work. It\nis offered as speculative: a hypothesis about a hypothesis.
What the frame notices is that the corpus has seen this shape before. The Wheel\nof Heaven reading of \nAbraham\n does not treat\nhim as the founder of a new faith. It treats him as a figure recruited after\nthe Sodom intervention — the tested, verified leader around whom a diminished\nlineage was reorganised, the founder of a recovery program. The corpus states,\nin the Abraham entry, that the Abrahamic lineage \"continues to produce\nalliance-cultivated traditions across subsequent ages.\" The pattern is\nrestoration: each major tradition reaching back to re-establish an original it\nbelieves the previous custodians corrupted.
\nThe Hanafiyya is that pattern naming itself out loud. A movement whose own\nearliest self-description is \"the recovery of the religion of Abraham,\" set\naround an altar attributed to Abraham and Ishmael, transmitted through the line\nthe \nAbraham\n entry already flags as the branch\nthe alliance \"protects\" and invests in alongside the covenant line — this is not\na new datum the corpus has to strain to fit. It is the corpus's own thesis\nappearing in seventh-century dress.
\nThree features sharpen the fit, and a Petra setting tightens each:
\nThe first is restoration over novelty. The corpus reads the great\ntraditions as successive cultivations of one lineage, not as unrelated\ninventions. The Hanafiyya states that logic from the inside — older than Judaism\nand Christianity, recovered rather than founded.
\nThe second is the operational lexicon carried by language. The\n\nMuhammad\n entry already notes that the Qur'an\npreserves vocabulary cognate with the Hebrew operational terms the corpus\ntracks: malak beside mal'akh (messenger), rūḥ beside ruaḥ (spirit,\nbreath), sakīna beside shekhinah. Those cognates run through\nAramaic.[10]\n If the Qur'an's language descends from\nNabataean Aramaic, then the new scripture inherited the older record's\noperational words through the same channel that carried its alphabet — the\n\nNabataeans\n, an Arab people literate in the\nlanguage of \nJesus\n and of Second-Temple Jewry.\nThe restoration the Hanafiyya claims at the level of religion is mirrored at the\nlevel of vocabulary.
\nThe third is editorial layering, a theme the corpus returns to across the\nprophetic record: the gap between an originating event and the canonical record\nthat later fixes it. Gibson's al-Ḥajjāj — who, on this reconstruction, altered\nthe qibla, re-pointed the text, and consolidated the move south, all within the\nfirst Islamic century — is a concrete instance of exactly that gap. The corpus\ndoes not need Gibson's specific reconstruction to be true in every detail to\nrecognise the shape. It is the same shape the corpus reads in the editorial\nhistory of the Hebrew and Christian canons: an original overwritten by its\ncustodians, recoverable only by reading the record against itself.
\nThis is the \nframework's\n existing reading of\nMuhammad, undisturbed. The corpus places his career in the early-to-middle\nAge of Pisces — its period of broad cultural\ntransmission of the operational message across civilisational regions. A sanctuary tradition relocated and\nreformulated within a single century is consistent with that picture of\ntraditions edited in transmission. The Petra hypothesis, read through the frame,\ndoes not overturn the corpus's Muhammad. It supplies a geography for the\nHanafiyya the corpus had already named.
\nThere is a temptation to read the ḥanīf's turn as the familiar story of\nmonotheism replacing polytheism — many gods collapsing into one. The Wheel of\nHeaven frame declines that reading, and the reason cuts to the centre of the\ncorpus. The \nElohim\n are a plurality: a\nCouncil of finite, embodied creators, the\ngrammatically plural Elohim the Hebrew text keeps and that the corpus reads\nacross the \nwhole record\n. The\ncorpus stands, if anything, closer to a pantheon of real and locatable beings\nthan to the abstract, omnipresent, omniscient, omnipotent single God of\ndeveloped theology — a God the Raëlian source material names as a translation\nerror, the moment the plural creators were \"transform[ed] into a single\nincomprehensible God.\" A frame that does not hold the truth to be metaphysically\none cannot read the Hanafiyya as the triumph of metaphysical oneness.
\nThe contrast the frame does read is between idolatry and knowing the\ncreators — and the canon states it directly, in the same post-Sodom collapse\nthat frames Abraham as the recovery figure:
\nBut men, having fallen back into a very primitive state after the destruction of\nthe most intelligent and of centers of progress like Sodom and Gomorrah, began\nstupidly to worship pieces of stone and idols, forgetting who had created them.\n
The fault named here is not believing in too many beings. It is worshipping wood\nand stone and gold — dead statues — having forgotten who actually made\nhumanity. And the remedy is not the dissolution of the divine into a single\nabstract point; it is the recovery of accurate knowledge of the real makers. The\nfour seekers' verdict on the idol they had circled — that it \"could neither\nhear, nor see, nor hurt, nor help\" — is a judgment on lifeless images, not on\nplurality. Read through the frame, the \nHanafiyya\n's\n\"oneness\" is the singleness of the truth about the creators recovered against\nthe scatter of meaningless idols that had buried it — not the philosophers'\nOne. The turning the ḥunafāʾ sought was a turning back toward the Elohim, the\nactual gods, whom their people had forgotten. This is the corpus's\ninterpretation laid over the canon's own premises, offered as such; the\nElohim-plurality and the idolatry-as-forgetting are stated in the Raëlian\nmaterial, while their fit to the ḥanīf/idolater distinction is the reading.
\nNone of this is owed belief, and the editing standards of this site require the\nstrongest objection to be stated in its own voice rather than caricatured.
\nThe most serious opponent is David A. King, the leading historian of\nmedieval Islamic astronomy, and his objection is not that the early mosques face\nMecca. He grants that many early mosques do not face the Hijazi Mecca in the\nstrict geographical sense. His objection is to the inference that they were\ntherefore aiming at Petra. King's alternative is the folk-astronomical\n\"sacred geography\"[12]\n preserved in medieval Islamic texts:\nearly Muslims, lacking spherical trigonometry, oriented their mosques toward the\nKa'ba by non-mathematical schemes — aligning with the rising or setting of\nparticular stars, with the cardinal directions, with the winds associated with\nthe Ka'ba's own walls. On King's account the early qiblas are a patchwork of\nthese local schemes, and the apparent convergence on Petra is an artefact of\nGibson's method rather than a fact about the builders' intentions.
\nThe exchange is genuine and unresolved. King's critique runs to a\nmonograph-length treatment; Gibson devotes a chapter of Let the Stones Speak\nto answering it, and Schumm and Goldstein add a statistical appendix. Gibson's\ncounter is that King's model multiplies alignment-schemes until it can\naccommodate almost any orientation — that a theory able to fit every mosque\npredicts none of them — and that it carries an uncomfortable implication: that\nfor three centuries Muslims were unable to determine the direction of their own\nsanctuary and merely approximated it many different ways. A reader who finds\nKing's many schemes more historically plausible than a single relocated target\nwill not be moved by the qibla data, and should not pretend to be. The honest\nstatement is that the central archaeological claim is live, not settled.
\nSeveral further cautions belong on the record. The Macoraba argument is\nnegative, and the absence of Mecca from Ptolemy is weaker evidence than the\npresence of Petra in the qibla would be. The Ka'ba-measurement match rests on an\nunexcavated structure. Luxenberg's Syro-Aramaic readings are rejected by much of\nthe field. And the whole hypothesis sits outside mainstream Islamic-studies\nconsensus, which continues to locate Islam's origins in the Hijaz, and which\nreads the same gap between event and source — the \"200 silent years\" — without\nconcluding that the holy city moved. It is distinct from, and more specific\nthan, the older source-skepticism of Crone and Cook's Hagarism; it shares that\nschool's willingness to read the material record against the later literary\nsources, and inherits that school's burden of proof. Fred Donner's reconstruction\nof an early, ecumenical \"community of believers\" reaches some convergent\nconclusions about Islam's gradual self-definition without relocating its\ngeography at all — a reminder that the name argument and the map argument can\nbe separated, and that one can accept the Hanafiyya thesis while suspending\njudgment on Petra.
\nThe strength of the Petra hypothesis is not in any single strand, each of which\na specialist can contest. It is in the convergence — the way an archaeologist\ncounting qibla walls, a statistician testing scatter, a cartographer\nre-projecting Ptolemy, a historian reading al-Azraqī's tape measure, and a\nlinguist tracing the al- article all end up pointing at the same abandoned\ncity in southern Jordan, for reasons that have nothing to do with one another.\nConvergence of independent lines is how circumstantial cases are built, and it\nis also how coincidences are mistaken for cases. Which one this is will be\ndecided by excavation at Petra, by the scholarly response to King, and by\nwhether the qibla dataset survives independent re-measurement.
\nThe Wheel of Heaven corpus does not need that verdict to draw its own, narrower\nconclusion. The part of this that rests on the Qur'an and Ibn Isḥāq — that the\nnew religion understood itself as the \nḤanīfiyyah\n,\nthe recovered religion of \nAbraham\n, through the\nline of Hagar and Ishmael — stands whether or not the holy city was Petra. And\nthat part is the part the corpus was already prepared for. The framework reads\nthe Abrahamic lineage as a recovery program that keeps producing restoration\ntraditions across the ages. In the Hanafiyya, one of those traditions states the\nprogram in its own words. Gibson's stones, if they speak as he hears them, would\ntell us where it stood. The texts already tell us what it thought it was: not a\nbeginning, but a return.
\nWhat follows is mine — the author's conviction, set apart from the argument\nabove, which stands or falls on its own evidence.
\nIn September 2023 I drove from Jerusalem to Petra and back, and spent a few\nnights near the site, walking it thoroughly. I already knew Gibson's work, and\nmoving through those gorges with it in mind was one of the more enthralling\nthings I have done — reading the monuments as a candidate holy city rather than\nas a ruin on a postcard. To me it simply coheres. Petra sits close to Jerusalem,\nclose to the country where Abraham and Ishmael lived; the Nabataeans and their\nAramaic-into-Arabic script are right there; and it makes a quiet, natural sense\nthat this is where Muhammad came from — in an age when the land was still green,\nperhaps at the very beginning of the desertification that would later swallow so\nmuch of Arabia. Whether that drying was something the Elohim anticipated I cannot\nsay; I note it only as a thread worth pulling.
\nMy fascination with Gibson sits beside something the Raëlian source material\ngives me: the recognition of Islam as a legitimate link in the prophetic chain,\na genuine Elohimic religion rather than a rival to be explained away. Holding\nboth at once leads me to a conclusion the corpus already reaches for elsewhere —\nthat Islam was corrupted over time, as Judaism and Christianity were, and as very\nnearly every religious and mystical tradition seems to be. It may be human\nnature: take something Elohimic and, within a few generations, the custodians\nturn it into a geopolitical instrument, a contest for land and power, and the\noriginal intent silts over. Islam, in its later imperial career, became a heavy\ninstance of that drift — which is no argument against what it began as, only\nagainst what was done to it.
\nThat is why I want, when there is time, to re-translate the Qur'an through the\nWheel of Heaven lens — and after it the Hadith and other texts — the way the\nTranslation Program has already begun with Genesis, Exodus, and the\nrest. The goal is not to make the Qur'an say something new. It is to read it, as\nfar as the language allows, the way its first reciters may have meant it: as the\nrecovered religion of Abraham, pointed back toward the creators their people had\nforgotten. — Zara Zinsfuss
\n", "extra": {"article_type":"explainer","author":"Zara Zinsfuss","author_slug":"zara-zinsfuss","category":"Comparative","claim_type":"speculative","editorial_pass":"2026-05","footnotes":[{"content":"Arabic *qibla*: the direction a Muslim faces in prayer, marked in a mosque by the *miḥrāb*, a niche set into the *qibla* wall. Since the standardisation of the rite it has meant the direction of the Ka'ba; the question this essay turns on is whether the *earliest* mosques were in fact built to face it."},{"content":"Anno Hegirae, 'in the year of the Hijra' — the Islamic calendar, counted from Muhammad's emigration from his home city to Yathrib (Medina) in 622 CE. 1 AH = 622 CE; 100 AH falls c. 718 CE. The lunar Islamic year is about eleven days shorter than the solar one, so the two calendars drift apart over centuries."},{"content":"Al-Ḥajjāj ibn Yūsuf al-Thaqafī (d. 714 CE), the formidable Umayyad governor of Iraq and the eastern empire under ʿAbd al-Malik and al-Walīd. Tradition already credits him with adding the vowel-pointing to the Qur'anic consonantal text; Gibson's reconstruction casts him as the central 'reformer' who also altered the qibla and consolidated the move south."},{"content":"The *rasm* is the bare consonantal skeleton of an Arabic text, without the dots that distinguish letters of the same shape and without vowel marks. Early Qur'anic manuscripts are written in *rasm*; the same skeleton can often be vocalised in more than one way, which is what makes the Aramaic-substrate arguments possible."},{"content":"A *baetyl* (Greek *baitylos*) is a sacred standing stone worshipped as the seat or embodiment of a deity. Nabataean religion was rich in them, typically plain cubic or rectangular blocks rather than carved images — a habit of stone-veneration in cubic form that the Petra hypothesis sets beside the cube of the Ka'ba and its Black Stone."},{"content":"Claudius Ptolemy's *Geography* (c. 150 CE) is a gazetteer of some eight thousand places given as latitude–longitude pairs, compiled in Alexandria. It survives not as a map but as the coordinate list, from which maps are reconstructed. Ptolemy systematically underestimated the size of the deserts of Arabia, so his coordinates there must be re-projected against known control points before any modern identification can be trusted."},{"content":"The *Sīra* is the biographical-tradition genre on Muhammad's life. The earliest, Ibn Isḥāq's *Sīrat Rasūl Allāh* (c. 760 CE), survives mainly through the later recension of Ibn Hishām (d. 833 CE). It is the principal narrative source for the pre-revelation period, including the four seekers of the Hanafiyya."},{"content":"Arabic *ḥanīf* (plural *ḥunafāʾ*): one who has turned away from idolatry to the upright, primordial monotheism. The Qur'an applies it to Abraham repeatedly. The abstract noun *Ḥanīfiyyah* names that religion. The term is unrelated to the later Sunni legal school (the Ḥanafī *madhhab*) named for the jurist Abū Ḥanīfa, despite the shared root."},{"content":"The Nabataeans were an Arabic-speaking people who wrote in Aramaic; their cursive Aramaic hand is the direct ancestor of the Arabic alphabet. This script-descent is uncontested mainstream epigraphy, independent of any thesis about where Islam began."},{"content":"A *calque* or loan-translation renders a foreign phrase part by part into the borrowing language rather than importing the word itself. Several Qur'anic terms of the operational lexicon — *malak*, *rūḥ*, *sakīna* — are cognate, through Aramaic, with the Hebrew *mal'akh*, *ruaḥ*, *shekhinah* the Wheel of Heaven corpus tracks across the older record."},{"content":"ʿAbd Allāh ibn al-Zubayr led a rival caliphate (683–692 CE) from the holy city against the Umayyads of Damascus, and rebuilt the Ka'ba during his reign. His defeat by al-Ḥajjāj is one of the fixed points of the period; Gibson relocates the siege from the Hijaz to Petra, citing the catapult-stone deposits excavated at the Great Temple."},{"content":"The folk-astronomical 'sacred geography' of David A. King: a body of medieval Islamic texts that assign mosque orientations to the rising and setting points of stars, to the cardinal winds, and to segments of the Ka'ba's own perimeter — non-mathematical schemes for facing the sanctuary from a distance, which King argues account for the early orientations without any aiming at Petra."}],"keywords":["Petra","Mecca","qibla","Dan Gibson","Hanafiyya","Nabataeans","religion of Abraham","Qur'anic Arabic","Islamic origins","Ishmael"],"lang":"en","lang_prefix":"","references":[{"id":"the-book-which-tells-the-truth","locator":"'The Truth' chapter, 'The Sacrifice of Abraham' (Abraham as tested, recruited figure of the recovery)"},{"id":"let-the-stones-speak","locator":"The principal statement of the qibla, Mecca-antiquity, Ka'ba-measurement, and name-migration arguments"},{"id":"qur-anic-geography","locator":"The four-rivers reconstruction of Ptolemy's Arabia; the geographical case"},{"id":"early-islamic-qiblas","locator":"The underlying mosque-orientation dataset (also the online Qibla Tool)"},{"id":"qibla-accuracy-schumm-goldstein","locator":"Statistical test: early qiblas consistent with intentional Petra orientation, error growing with distance"},{"id":"from-petra-back-to-makka","locator":"King's folk-astronomical 'sacred geography' counter-model — the principal scholarly rejection"},{"id":"mecca-and-macoraba","locator":"Al-ʿUṣūr al-Wusṭā 26 (2018): no secure evidence for Mecca's antiquity; the Macoraba identification dismantled"},{"id":"on-the-origin-of-quranic-arabic","locator":"Qur'anic Arabic descends from Nabataean Arabic; the al- article and rasm evidence"},{"id":"graeco-arabica-southern-levant","locator":"Epigraphic basis: the Aramaic script of the Nabataeans 'casts a clear Arabic shadow'"},{"id":"the-life-of-muhammad-ibn-ishaq","locator":"Guillaume's translation, pp. 98–103: the four seekers of the Hanafiyya"},{"id":"akhbar-makka","locator":"The irregular four-sided Ka'ba measurements"},{"id":"the-history-of-al-tabari","locator":"Vols. XXI–XXIII (the Marwanid civil war): Ibn al-Zubayr's siege, al-Ḥajjāj, the Ka'ba rebuilding"},{"id":"the-syro-aramaic-reading-of-the-koran","locator":"The contested Syro-Aramaic substrate strand"},{"id":"hagarism-the-making-of-the-islamic-world","locator":"The maximalist source-critical revisionism from which Gibson's archaeological argument is distinguished"},{"id":"muhammad-and-the-believers","locator":"The 'community of believers' reading: Islam's gradual self-definition without relocating its geography"},{"id":"the-qur-an","locator":"3:67; 2:135, 142–145; 6:92; 14:4; 16:123 (the Hanafiyya / millat Ibrāhīm verses and the qibla-change verses)"}],"summary":"For roughly the first Islamic century, the mosques do not face Mecca. On Dan Gibson's survey of the oldest datable sanctuaries, their qibla walls point north — to Petra, the abandoned Nabataean capital in southern Jordan. That single archaeological claim sits at the centre of a convergence: Mecca is absent from Ptolemy's map of Arabia; the earliest Ka'ba measurements describe an irregular four-sided structure that matches an altar in Petra, not a cube; the grammar and script of the Qur'an descend from Nabataean Aramaic, not from any Hijazi Bedouin dialect; and the earliest sources call the new religion not 'Islam' but the *Ḥanīfiyyah* — 'the religion of Abraham.' This Explainer works through each strand in turn, gives the principal critic (David A. King) his own voice, and then reads the whole through the Wheel of Heaven frame: as one more turn of the Abrahamic-restoration cycle the corpus already tracks — a movement that understood itself, explicitly, as the recovery of a lost original religion through the line of Hagar and Ishmael."} }, "links": { "self": "/v1/articles/the-first-mosques-faced-petra-not-mecca/", "canonical_html": "https://www.wheelofheaven.world/articles/the-first-mosques-faced-petra-not-mecca/", "index": "/v1/articles/", "schema": "/v1/schema/article/", "related": [] } }