Arrivals at the Temple of the Kyber: A Reconstruction of Pilgrimage to Jedha from Surviving Register Copies, 40–26 BBY

1. Introduction

Pilgrimage to Jedha is easier to assert than to measure. The moon is a desert world regarded as holy by many faiths, and its holy city, NiJedha, holds the Temple of the Kyber, a site associated in the traditions we have read with the Force and with the kyber crystals, which those traditions treat as attuned to it. The temple is kept by the Guardians of the Whills, and travellers reach it from across the galaxy. Beyond those points the record thins quickly. Comparative studies of pilgrimage have leaned on travellers' narratives and on the reputation of the site, both of which favour the memorable journey over the ordinary one (Ostrelle, 33 BBY).

This article asks a plainer question: how many parties came, when, from where, and with what observances? Our answer rests on one class of evidence, the temple's own pilgrim registers, as preserved in copies at the Royal Alderaanian Academy of Sciences (Temple of the Kyber, 40–26 BBY). The vantage of the study is 25 BBY, and the reconstruction covers the standard years 40 to 26 BBY. We aim for humble precision. Every annual figure carries a range, and every claim about the pattern of arrivals is qualified by what the copies can and cannot show.

We make one further limitation explicit at the outset. The registers record acts of arrival and offering. They are not statements of belief, and any theology we attribute to pilgrims is our own reading of what they chose to write down. We say nothing of the doctrine of the temple's keepers, on which the copies are silent.

2. Sources and Method

The registers are running lists of entries in a single temple hand, each entry naming a party or an individual, a declared place of origin where one was given, a date in the temple's own reckoning, and a record of what was offered or observed. The academy holds copies of the gatherings for thirteen standard years. No copy of the gatherings for 34 BBY or 29 BBY was ever delivered, and we cannot say whether they were never made or were later lost. The copies were made by academy scribes in several campaigns, and a collation and condition survey of the holdings was published in 28 BBY (Royal Alderaanian Academy of Sciences, 28 BBY). We follow its foliation throughout.

Two source-critical problems dominate. The first is loss. Each year's gathering carries continuous folio numbers, so the fraction of a year that survives can be estimated as legible folios over foliated folios, on the assumption that entries per folio are roughly constant within a year. Six years survive at a legible fraction of 0.94 or better. Others are damaged, and the worst, 36 BBY and 26 BBY, retain about half. For each year we give a point fraction and a range that reflects folios only partly legible and gatherings whose closing folio number is itself lost. Reconstructed arrivals are legible entries divided by the fraction, with the range taken from the extreme fractions.

The second problem is copying error. Scribes copying long lists make characteristic mistakes, among them repeated lines at the joins between gatherings, misread numerals and normalised place names (Hadrevane, 27 BBY). We removed 41 apparent duplicates, which leaves 3,611 legible entries across the thirteen years, and flagged 96 entries whose date or origin was doubtful. Flagged entries are retained in the annual counts; a doubtful origin is coded as unstated and a doubtful date as undated. Entries record parties, so we count entries and make no claim about persons. Parties travelling together are not independent observations, a point we return to in the analyses that follow.

Two variables were coded by hand: declared region of origin (Core, Mid or Outer regions, or unstated) and the principal offering or observance. We combine Mid and Outer regions because the registers name origins in local terms that we could not sort into the two with confidence. A subsample of 240 entries was coded twice, independently. Agreement on region was 92.5% (kappa 0.87), and on offering type 85.0% (kappa 0.78). Offering type could be coded for 3,047 of the 3,611 entries, or 84.4%.

3. Annual Arrivals and Their Rhythm

Table 1 gives the reconstruction. Point estimates run from 280 recorded parties in 26 BBY to 392 in 32 BBY, and the thirteen-year mean of the point estimates is about 341. The ranges are wide where the fraction is poorly known. The figure for 36 BBY, at 374, could lie anywhere between 312 and 468, so that year alone cannot be called a high or a low. Only the six best-preserved years give ranges no wider than about forty-five entries, and it is on these that we place most weight.

The series shows no detectable trend. A log-linear fit of the point estimates on year, with an indicator for 32 and 31 BBY described below, gives a change of +0.6% per standard year going back in time (95% interval, from a fall of 0.8% to a rise of 2.0%, p = .38). The data are compatible with a temple whose traffic held level for fifteen years, and equally with modest drift in either direction. The loss of two whole years, and of half of two others, means that a short-lived surge or slump in the missing gatherings would be invisible to us.

Seasonal rhythm is visible, though faint. Of the 3,611 entries, 2,564 (71.0%) could be placed in a quarter of the standard year by our concordance between the temple's reckoning and the standard calendar. The concordance is our own construction from dated entries that name recognisable observances, and a mistake in its anchor would rotate the seasons but not remove the pattern. The first quarter holds 702 entries (27.4%), the second 617 (24.1%), the third 588 (22.9%) and the fourth 657 (25.6%). The departure from an even split is significant on a chi-square test with three degrees of freedom (11.49, p = .009), but the test treats entries as independent and parties are not. We read it as descriptive. The peak-to-trough ratio is only 1.19.

We can offer one hypothesis for the pattern and no more. Sources on the care of arid-world sites of pilgrimage describe seasons of heavy wind and dust that make approach by land and by repulsorlift craft harder (Tholwyn, 31 BBY). If Jedha has such a season, and our sources for the moon's own climate are thin, a third-quarter trough would follow. The registers do not record weather, so the hypothesis cannot be tested here.

4. Origins and Observances

Declared origin changes over the period. In the six early years, 40 to 35 BBY, the 1,603 entries divide into 593 from Core regions (37.0%), 657 from Mid or Outer regions (41.0%) and 353 with no origin stated (22.0%). In 33 to 26 BBY the 2,008 entries divide into 904 Core (45.0%), 743 Mid or Outer (37.0%) and 361 unstated (18.0%). The two distributions differ (chi-square, two degrees of freedom, 24.87, p < .001), though the test treats entries as independent and parties are not, so it too is descriptive. Over the whole record 41.5% of entries name a Core origin, 38.8% a Mid or Outer origin and 19.8% none.

Two cautions apply. The fall in unstated origins may reflect a change in scribal practice, a fuller form of entry or a stricter keeper, and not any change among pilgrims; the registers themselves are our only guide, and the copies do not reproduce the ruling of the columns. Nor is the shift tied to any external event. The later period includes the blockade years, and 34 BBY is missing, so the pooled comparison cannot show whether the shift began before the blockade or after it; the two periods also differ in many ways besides origin. The pattern is compatible with a temple whose reputation travelled more widely among wealthy and well-connected worlds, but that is a hypothesis about audiences, and the registers do not test it (Marrowen, 30 BBY).

The record of offerings is more uniform. Of the 3,047 entries with a codable offering, 1,204 (39.5%) record lamps or other light offerings, 812 (26.6%) record a vigil or observance with no object, 683 (22.4%) record an inscribed petition, and 348 (11.4%) record a material gift left with the temple. Light offerings therefore outnumber every other kind, and vigils without objects are more common than petitions. The proportions do not sum to exactly 100% because of rounding.

Petitions are the closest the registers come to the pilgrims' own voice, and we read them with care. Their surviving formulas most often address the Force directly, and a smaller number invoke the crystals as a focus or a witness for the petitioner's intention. Requests for safe passage and for the healing of a named relative recur. We did not code petition content in a way that would support figures, and what follows is an impression from reading, not a finding. It does not amount to a doctrine of the temple. Pilgrims of many faiths are said to visit Jedha, and the petitions may equally reflect the habits of the petitioners and the scribes, or the phrases of a familiar form. The record gives no means of distinguishing these.

A tradition preserved in the Jedi Archives holds that members of the Jedi Order in earlier ages travelled to the temple in connection with the crystals, and the Archives have released a catalogue note to that effect in abridged form (Jedi Archives, 30 BBY). We report it as a tradition and we cannot confirm it. The Order's gathering of crystals is separately attested at Ilum, and nothing here bears on that practice. No entry in the copies declares membership of the Order, but the registers seldom record affiliation of any kind, and an absence in such a source is weak evidence.

5. Arrivals After the Blockade of Naboo

The blockade of Naboo in 32 BBY was a public event, debated in the Galactic Senate and known across the Republic (Galactic Senate, 32 BBY). We asked whether recorded arrivals at the temple rose in its aftermath. The window is a post-hoc choice of ours: 32 and 31 BBY against the seven earlier years with data, 40 to 33 BBY. The event fell within 32 BBY, so that year straddles it, and any p-value is conditional on the window chosen. A different window would give a different figure.

On the point estimates the mean in the window is about 386 parties a year, against about 335 before it, a ratio of 1.15. With the trend term included, the multiplier attached to the window is 1.18 (95% interval 0.99–1.40; t on 10 degrees of freedom 2.11, p = .061). A simulation that redraws each year's fraction across its stated range gives a median multiplier of 1.18 with a central 95% interval of 1.10 to 1.26. That interval reflects uncertainty in the losses alone. It is narrower than the regression interval because it ignores year-to-year variation, and the regression interval is the fairer statement of what thirteen annual points can show. The regression is unweighted and its interval takes no account of error in the reconstruction itself. The window years are among the better preserved, with fractions of 0.95 and 0.92, so the result does not come from heavy correction of thin years. As a sensitivity check, 31 BBY alone as the window (32 BBY set aside) gives a multiplier of 1.16 (95% interval 0.90–1.49, p = .21), and 32 BBY alone (31 BBY set aside) gives 1.19 (0.93–1.53, p = .14). Both point the same way as the two-year window, and neither reaches conventional significance.

Taken together, the result is a small sample with a window whose length is our choice and a p-value that just misses the conventional threshold. It is suggestive of a rise of some tens of parties a year, and nothing stronger. The registers do not tell us why any party came. Several accounts are possible: travellers seeking observance in a period of public anxiety, a displacement of routes, or an ordinary fluctuation of the kind the series shows elsewhere. The estimate for 36 BBY alone spans 156 parties, several times the excess claimed for the window.

6. Discussion

The reconstruction supports four statements. The temple received a fairly steady flow of recorded parties, with a point-estimate mean near 341 a year and no trend we could detect. Arrivals were slightly more frequent in the first quarter of the standard year than in the third. The declared origins of pilgrims moved toward the Core between the earlier and later halves of the record. And recorded arrivals in the two years after the blockade of Naboo were higher than before, by an amount that the data neither establish nor exclude.

Each statement is narrower than it might appear. The flow is steady only across the thirteen years we can see, and 34 BBY and 29 BBY are missing altogether. The seasonal pattern rests on the 71.0% of entries that could be dated, through a concordance of our own devising, and on a test that overstates independence. The origin shift may be partly a change in what scribes wrote. Only the blockade comparison invites a causal story, and it is the weakest of the four.

A larger lesson concerns method. A pilgrimage site is easily described through its most eloquent visitors, and the registers show how different the ordinary record is: lamps far more often than gifts, a vigil slightly more often than a petition, and a steady rather than a spectacular rate of arrival. Those observances fit a site of many faiths, since a lamp or a silent vigil requires no shared creed. This is an interpretation of the pattern of offerings and not a finding about belief. What the temple's keepers made of these visitors lies outside the registers, and we have not attempted to reconstruct it.

7. Limitations of the Record

The evidence is a set of copies, made at intervals by scribes whose fidelity we can assess only from internal evidence. Where a copyist normalised a place name or silently dropped a line, the loss is invisible to us. The fractions used to correct for loss assume a constant number of entries per folio within a year, which is unlikely to hold across gatherings written in different hands. The correction is therefore an approximation, and the stated ranges do not include the assumption itself.

Coverage is uneven in time. Two years are absent, the worst-preserved years include 36 BBY and 26 BBY, and 40 BBY is also thinly covered, so both ends of the record and its early middle are weak points for detecting a trend. The origin analysis merges Mid and Outer regions and treats 96 flagged entries as unstated or undated. The seasonal analysis uses a concordance that has not been checked against an independent calendar of the temple. Petition content was read but not coded, and reliability was assessed only for region and offering type.

Finally, the registers describe those who signed them. Pilgrims who came and left no entry, whether through custom, haste or exclusion by the temple's keepers, are absent, and there is no way to estimate their number. The counts are counts of recorded parties, and are best used to compare the years of the record with one another, not to estimate the total traffic to Jedha.