Abu Kamil's algebra with irrational coefficients

Abu Kamil Shuja ibn Aslam did more than obtain a radical at the end of a solution. Cengiz Aydin explains that his algebra admitted irrational quantities into equation coefficients as well as solutions. This enlarged what algebra could handle; it was not invention of all algebra or discovery of the first length that cannot be written as a ratio of integers ¹.

We record a refinement within an inherited tradition. Melek Dosay presents radical problems including a product containing square roots of3 and2. We read that example on an actual image of the modern paper, not in Abu Kamil's manuscript. The level of evidence must remain visible ².

A coefficient is a quantity multiplying an unknown or one of its powers: it may be2, one half, or in this topic a quantity such as the square root of2. “Root” itself needs care: a square root of a number and a solution of an equation are related but different meanings. An equation having an irrational solution is not the same as an equation having an irrational coefficient. Aydin distinguishes these positions when describing the refinement ¹.

In modern teaching notation, x²=2 has integer coefficients and an irrational positive solution. In x²+sqrt(2)x=4, an irrational quantity appears in the coefficient itself. This second equation is our teaching example, not a purported line from the book. The difference concerns the problem's structure before its answer, not the length of the printed symbol.

MacTutor records Abu Kamil acknowledging al-Khwarizmi's algebraic authority; Aydin connects his work with earlier algebra and Euclidean geometry. The maker's older record also preserves the1939 Cairo editors' references to that acknowledgement. The story does not begin in a world without unknown quantities or geometry without irrational magnitudes ^{3 1 4}.

Radicals in geometry predate this application, but making a geometric quantity act as a factor in algebra raises a different question from drawing it. We neither reduce all predecessors to one person nor assume that every Greek or Arabic inheritance reached the author by an identical route.

Very little personal information survives. MacTutor gives an Egyptian setting and approximate life dates850–930; TDV cautions about modern career claims and places him after al-Khwarizmi and before al-Imrani's commentary on his book. We therefore do not turn a late suggestion into a confirmed shipbuilding job or court appointment ^{3 5}.

The record places the work in a tenth-century activity context, not a known invention day. We have not read a manuscript colophon dating composition and cannot turn a copy date or modern edition date into the year these equations were developed. The maker's existing approximate floruit remains unchanged, rather than being adjusted to make the inventions list easier to sort. Missing an exact day does not prevent explaining a documented contribution; it prevents inventing precision.

O'Connor and Robertson's University of St Andrews biography explicitly names Abu Kamil as an Islamic mathematician. This is an individually recorded scholarly classification, not an inference from Arabic name, country or period. Sunni is separately the museum's declared frame for the Abbasid Egyptian setting; the biographies read attest no contrary tradition and do not supply a personal legal-school declaration ^{6 5}.

His role here is “improver,” while al-Khwarizmi appears as a named predecessor. Later scholars and translators do not thereby become coauthors of the original. Distinguishing contributor, transmitter, interpreter and modern editor clarifies rather than erases the work's history.

Dosay p66 presents, in modern symbolic notation, (x+sqrt(3))(x+sqrt(2))=20. We inspected the rendered page because text extraction omitted many of the article's formulas. This is not a transcription of an old Arabic line, nor evidence that an original manuscript used our letterx or radical symbol ².

Our modern expansion gives x²+(sqrt(3)+sqrt(2))x+sqrt(6)=20. An irrational quantity appears in the coefficient ofx and another in the constant term. They are not temporary marks waiting to collapse into integers: calculation must continue with them accurately. This expansion checks the published problem; it is not a quotation of Abu Kamil's method or a solution we claim Dosay supplied.

Our simpler example, x²+sqrt(2)x=4, can be checked by substituting the square root of2 forx: its square is2 and its product with the square root of2 is2, giving4. This does not claim to find every solution or solve all equations of every degree. It demonstrates an irrational coefficient and an irrational solution in one example.

One cannot add radicals merely by adding their radicands: the square root of2 plus the square root of3 is not the square root of5, because squaring the sum introduces 2sqrt(6). Multiplying these two positive radicals does give the square root of6. These are modern explanations we verify, not rules purportedly copied from an uninspected folio.

Exact and approximate values must also remain distinct. Decimal replacements can help with an approximate check, but rounded output is not the exact expression. Modern calculator digits do not become Abu Kamil's documented precision, and an algebraic solution does not automatically qualify as a positive geometric length.

To read the example in stages, distribute the multiplication, collect the two terms containing the unknown, and then inspect the resulting coefficient. The constant square root of6 does not vanish when the unknown terms are collected; moving it across the equality changes its sign. This lets a reader check each equality against the starting expression without attributing our school notation or sequence to a manuscript page. We are illustrating the meaning of the coefficient described by scholarship, not recovering a missing text ².

MacTutor and TDV connect Abu Kamil's work with al-Karaji and Fibonacci; TDV also describes commentaries and Hebrew and Latin translations. This is reception reported by scholarship, not an assertion that we inspected every witness or that every problem travelled unchanged ^{7 8}.

TDV distinguishes witnesses: a Latin translation containing the first two parts, a Hebrew version apparently incomplete relative to Arabic, and two Arabic witnesses described as complete and incomplete respectively. These are the entry's reports, not a collation performed here. An abridged translation and a complete book cannot be silently treated as identical reception evidence.

The book combines an algebraic tradition with Euclidean geometry. MacTutor describes three parts: quadratic equations, algebra applied to regular pentagons and decagons, and indeterminate problems. These do not all become one invention named “irrational coefficients”; they locate the refinement within a broader work ^{9 1}.

The biography also explains that higher powers were expressed verbally, not in our symbols. Usingx² or the abbreviationsqrt is an explicitly editorial convenience, not evidence that the author's terminology or page arrangement matched a modern school notebook ¹⁰.

Scholarship describes later algebraic reception and Fibonacci's use of the work, but assigning every contemporary formula to that influence requires a separate investigation. No direct edge leads here from this refinement to every symbolic-algebra program or design tool. Specific dependent records remain empty until dependence on this particular contribution is established ^{7 11}.

An absent automatic edge does not deny the reported reception. Narrative and graph address different levels: did a later author receive the book, and can a specific technical contribution be identified as transmitted? The first answer is not permission to invent the second.

The direct change is a broader range of computational and geometric problems that algebra can handle, not documented changes in wages or the speed of every shipyard. MacTutor describes a separate surveying book for government surveyors, offering rules and examples without proofs. Its audience cannot automatically be assigned to every part of the algebra ¹².

Today's visitor can benefit from distinguishing coefficient from answer, or exact expression from approximation. That is an educational value of our explanation, not a historical headcount of users or measured financial saving.

TDV describes a synthesis of practical algebra and theoretical geometric foundations, with expanded use of irrational quantities. Progress in algebra therefore need not mean merely naming an unknown or adding another exercise: it can mean changing the kinds of quantities on which calculations operate ⁸.

Calling this a step in algebraic history does not make Abu Kamil the founder of modern field theory or author of every current real-number rule. Those are later frameworks we must not place in his mouth. Appreciation concerns what the sources support, not what we can now write in a few symbols.

Aydin attributes an extension to irrational coefficients to Abu Kamil. We do not call him inventor of all algebra, the first radical or every operation involving geometric lengths. MacTutor preserves his acknowledgement of al-Khwarizmi: recognizing a later contribution need not erase its predecessor ^{1 3}.

Where scholarship uses “first” for a higher power or indeterminate problem, that is not a verdict transferred to everything on this page. Reading three accounts does not inventory every earlier text, and a modern radical sign cannot prove ancient symbolic priority.

Students can distinguish sqrt(2)+sqrt(3) from sqrt(5) and then expand a product that generates an irrational coefficient. The lesson does not require portraying older algebra as an incomplete computer: it had languages, purposes and reasoning methods described by the scholarship ¹.

The accompanying image is a modern right triangle released to the public domain by TheOtherJesse in2007. It explains sides and an angle, not Abu Kamil's drawing or an assumed isosceles example with sides1 and1. Its appearance supplies no historical proof about coefficients or the algebra book ¹³.

The directly read evidence in this run is MacTutor, TDV and specified Dosay pages, including the actual image ofp66. Arabic, Hebrew and Latin manuscripts and editions named in the entry were not inspected here. Their existence is reported with attribution, not an invented quotation or folio reference ^{1 2}.

Our modern arithmetic can be checked independently. A correct symbolic expansion today does not establish that every number in a modern paper is free of error or a literal reproduction of an older witness.

The story presents mathematics as dialogue between computation and geometry rather than a contest over names. The sources preserve Abu Kamil's acknowledgement of predecessors while explaining his additions, making continuity and change visible together. That is more informative than a headline announcing that he began everything ^{3 11}.

The formulas missing from extracted article text also show why digital research sometimes requires looking at the page. A blank in extraction is neither evidence that the source lacks an equation nor permission to fill it by guessing.

Draw a square and label its side “square root of2” and its area2. No cutting or exact measurement is necessary: these are symbolic cards. The square root of2 is not half of2; its square is2.

Now compare cards marked “square root of2” and “square root of3.” Combining the cards does not mean adding numbers inside the radical. If you know expansion, square their sum and find the additional product term. These are modern pencil activities, not manuscript reconstruction or a claim that the accompanying triangle depicts this square.

• O'Connor and Robertson, complete actual MacTutor biography ³.
• Cengiz Aydin, TDV1994,10:172–174, actual main text and author metadata ¹.
• Melek Dosay1990, selected pp57–60,66–68; p66 image inspected ¹⁴.
• Earlier source preserved in the maker record, not newly read in this run ⁴.
• Modern triangle metadata ¹³.