Al-Jazari's swordsman candle clock

A candle whose top does not simply sink as it shortens: instead it rises toward a fixed cap, periodically releasing a ball and moving a sword-bearing figure at the wick. This is the swordsman clock al-Jazari describes in his seventh clock chapter, using discharged balls to count elapsed night hours ². Its materials are a calibrated candle, counterweight, cords and ball channel, not flowing water or a pendulum. It is not the world's first candle clock: the author himself begins with a report of an earlier device whose mechanism he did not know ³.

A candle holder carries a long sheath, a falcon near its foot and a seated sword figure near its top. Inside are candle, supporting dish, weight, cords and a row of ball chambers. A falling ball links the falcon's signal to the figure's movement ². The exterior conceals mechanical relationships that an explanatory image partly opens to the reader.

This entry covers the complete chapter 7 assembly, not every candle clock in subsequent chapters or a separate invention for every ball, pulley and cap ¹. Calibration relates consumption of this candle to a reference duration; it does not mean drawing equal marks on any modern candle. The museum has not tested reconstructed hardware or its precision.

Al-Jazari says he had not read a candle-clock work or seen a completed example, but had heard of a candlestick with a lion's head releasing a ball each hour ³. The opening combines transmitted knowledge with a knowledge limit: appearance and result reached him, the mechanism did not. His personal experience cannot become a worldwide absence claim.

He says he considered making something like the reported device before describing his own arrangement ³. The earlier clock remains an external precursor in his account, without an invented maker or date. This is not a documented genealogy of every earlier or later clock, but an explicit engagement with an idea he had heard and sought to understand.

The account moves from the nighttime display to a uniform-section candle and measurement of its consumption during an independently measured hour. The working length is divided into fourteen periods, with two further divisions reserved in the arrangement ⁴. Preparing the timekeeping material is part of the design, rather than assuming every wax and wick burns at the same rate.

The author reports repeated arrangement, calculation and trials to perfect the ball-receiving pouch's movement ⁵. This is his account, not an experiment repeated here. TDV places his writing within Artuqid service at Amid. Hill's1204 or 1206 from manuscript differences remains alongside TDV's1205: a book-date disagreement, not a verified installation date for this clock ^{6 7}.

Al-Jazari is recorded as describer of an integrated improvement, not inventor of candles or all time measurement. His text explains relationships and acknowledges earlier hearsay; TDV identifies an engineer and mechanical author ^{2 6}. Unnamed craftspeople are not supplied from guesswork, and copyist and translator work is not reassigned to the original maker.

Sunni frame: TDV calls the named al-Jazari an Islamic scholar, the individual identity basis, not inference from name, court or origin ⁶. His biography's explicit Sunni default is the museum's frame here, not a new personal-school declaration. Farruq ibn Abd al-Latif is catalogued as calligrapher of the 1315 copy, not clock inventor or certainly the painter of every colour ⁸.

The candle rests in a dish connected by cords over pulleys to a weight. The descending weight pulls candle and dish upward, while the fixed cap permits advancement only as material disappears from the candle's top ⁹. The wick therefore remains near the cap instead of sinking with the shortening candle. This explains a relationship, not instructions for securing flame or preparing weights.

Alongside the candle is a channel divided into fourteen chambers, each holding a ball. Candle and dish prevent its escape; as they rise, another chamber level becomes exposed and its ball falls to the lower route ¹⁰. The ball does not originally measure the hour: calibrated candle consumption produces slow movement, then an uncovered chamber converts movement into a countable event.

The ball first reaches a suspended pouch whose cord is connected to the figure's arm. Its temporary load produces the described sword movement trimming the burned wick; the ball then travels into the falcon's head, its weight opening the beak before it falls to the base. Pouch and arm return ready for another signal ¹¹. This is a functional sequence, not a guide to making a bladed arm.

The system is limited by candle, balls and weight travel, not perpetual motion. The text provides fourteen signals and expressly limits its climatic application to the fourth region, rather than presenting a clock automatically suitable for every night everywhere ¹². Calibration against a reference period does not itself prove all subsequent periods match it. No independent timing-error measurements are available in this entry.

The evidence followed is a text and later pictures, not a list of working installations. TDV records manuscript copies, English translation and Arabic editing; Freer preserves a leaf from a dispersed copy dated December1315/Ramadan715 ^{6 8}. This establishes transmission of description and image, not that every copyist built working hardware.

The catalogue says the folio was probably produced in Syria; that qualification is not converted into a certain site of original clock manufacture ⁸. Its modern museum location likewise does not establish historical technical adoption there. Cities and installation years unsupported by the sources read remain absent from the spread list.

The assembly combines calibrated consumption, weight-and-cord lifting, a ball restrained then released, and transfer of its load into movement and visible representation ². Integration does not make every element new. The lion-head report also shows that linking a candle to a falling ball reached the author before this description, even though its particular mechanism remained unknown to him.

Comparison with the castle clock distinguishes timing materials: regulated water there, candle consumption and lifting here; both descriptions include preparation and calibration ¹³. This is a comparison between chapters, not proof that the candle clock derives directly from every water-clock detail or that all modern clocks descend from this particular linkage.

Hill draws a modern cutaway interpretation to expose mechanisms concealed by an exterior view, explicitly making the candle thicker than its real proportions ¹⁴. That image is not a ruler for measuring historical hardware. He also chooses one of two sheath lengths supplied by the text: an interpretive choice, not a measurement agreed across all witnesses.

He notes insufficient weight travel and excessive ball height in Oxford manuscript figure 74, comparing other drawings ¹⁵. His reading of the cap as an insertion-and-turn, or bayonet, fitting interprets wording he acknowledges as obscure; it does not establish worldwide priority for every similar locking joint ¹⁶.

The stated function is knowing elapsed night hours by counting balls, alongside the wick-trimming movement ¹⁷. It combines countable indication and mechanical spectacle, rather than a continuously graduated modern wristwatch face. Separating the signal from the timing material helps explain how discrete events can represent a continuous change.

That does not establish household ubiquity, low cost or regulation of all urban activity. The text specifies a candlestick, material and preparation, not a census of owners or public appointment records ². Its documented effect remains a function presented in an account; the social consequences of all later timekeeping developments are not automatically assigned to it.

The account makes calibration material rather than rhetorical: actual consumption during a reference period is related to a distance that exposes a ball chamber ⁴. Its historical interest is the described arrangement of consumable material, signal thresholds and represented action, not discovery or first measurement of time.

It also exposes boundaries between text, picture and interpretation. Treating the Oxford drawing as a perfect working plan would overlook Hill's weight-travel problem; treating his pedestal conjecture as archaeological fact would erase his inferential language ¹⁵. The entry retains these evidence levels rather than turning machinery history into an unbroken, unambiguous march of progress.

Author's account within the declared Sunni frame: Al-Jazari reports an earlier lion clock and his ignorance of its mechanism, then describes what he made and repeated movement trials ¹⁷. Credit concerns this detailed arrangement. Not having seen a completed example does not prove none existed anywhere, and the earlier report is not deleted to create absolute originality.

Academic interpretation: Hill praises attention to specification and calibration, but identifies drawing problems and infers pedestal details from other chapters ¹⁸. This is not a competing-inventor claim requiring a verdict; it is bounded attribution with differing certainty levels. The cap is not split off and counted again without independent priority research.

The clock can be studied as an analogue arrangement generating discrete signals: consumption and lifting progress gradually, while ball release is a countable event ². This is a teaching comparison, not identification as a digital computer or template for all software. Fourteen balls do not confer calendar calculation or resource-free self-setting.

No original machine or modern reconstruction was examined for this entry. No home experiment with flame, hot wax, lead weights or a sharp arm is proposed. Explaining the sword's textual function is not recommending imitation: a paper image can show restraint, release and counting without recreating physical hazards.

We read Hill's I.7 translation 83–86 and specific commentary 252–253, visually checking 83,86 and 253 in the shared reader ¹. These are actual selected readings, not full manuscript collation or inspection of every translation. The modern page read is not the public-domain picture reused here.

Freer F1930.71 preserves Arabic text and a clock illustration in a1315 folio credited to calligrapher Farruq ibn Abd al-Latif: paper, ink, pigments and gold, not a metal machine ⁸. A seated sword figure, candle, balls, cords, weight and falcon are visible, not al-Jazari's portrait. Reuse relies on Commons' faithful public-domain2D reproduction statement, not an assumed institution-wide image licence ¹⁹.

The spectacle joins a represented human, a falcon and a time signal. The text describes an enslaved character and supplies a colour description; these are historical context, not endorsement of enslavement or inference of a real person's origin from paint ³. The character is not evidence for the maker's identity, and the sword is not a battle narrative in this chapter.

The reported movement trims the wick and marks a period, rather than simply displaying a weapon shape ²⁰. The picture can attract attention before inviting a reader to follow cords and channels, provided decoration does not replace reading. Distinguishing figure from user and illustration from hardware is part of understanding material culture.

Use paper only: draw 14 circles for possible signals, then cross out one after each imaginary step. After 5 steps,9 remain:14 minus5 equals 9. At step 14 the counter ends, rather than continuing forever. The capacity comes from the account; steps and the choice of 5 are a modern teaching example ¹².

Draw a strip covering the circles, revealing one level per step without real balls, pulleys or weights. Equal game steps are not certification of historical precision. Make one story-step longer and notice that signal counting alone does not establish equal intervals. No wax, flame, blade or hot object is involved.

• Donald Hill's1974 translation, I.7,83–86 and specific commentary 252–253; actual reading and 83/86/253 inspection, not page reproduction.
• Sadettin Ökten, CEZERÎ, TDV, vol.7,1993,505–506; biography and individual identity basis.
• Freer F1930.71 and Commons metadata, institutional 1315 folio catalogue, inspected image and PD-Art rights.