1. Thirty Problems Each

In Venice in early 1535, two men exchanged thirty mathematical problems apiece. The questions were deposited with a notary, and a period was fixed for producing solutions. This was not an impromptu contest in mental arithmetic. Each competitor selected the ground on which he expected to prevail.

Antonio Maria Fiore possessed a rule for one type of cubic equation, learned from Scipione del Ferro, a teacher of arithmetic and geometry at Bologna. Del Ferro is conventionally credited with finding the method around 1515. The date is a historical reconstruction, not the date on a surviving autograph announcing the discovery.

Tartaglia's problems ranged more widely. According to his later account, he independently found the type of cubic solution on which Fiore relied and then solved all thirty questions rapidly, while Fiore solved none of his. The asymmetry matters more than the legend of two geniuses calculating at a table: one competitor had concentrated his advantage in a secret that the other learned to match.

The calendar needs care. Venetian years began on March 1, so February 1535 in modern year-numbering could be dated February 1534 locally. The original year number is one lower, not one higher. Copying a date without its calendar convention can shift an episode by a year.

The commonly repeated prize of thirty banquets is less secure than the exchange of questions, the notary, the deadline, and language about honor and rewards per successful solution. A picturesque prize should not be granted the status of the better-attested framework.

The central narrative source is Tartaglia's own publication eleven years later. He participated in the events and selected the version readers would receive. First-person evidence is valuable; interested first-person evidence still requires comparison.

Why had del Ferro not published the rule? No surviving self-explanation settles his motive. In the world of northern Italian mathematical teaching, rare problem-solving techniques could support reputation, patronage, and employment. That gives secrecy a plausible professional logic without proving what del Ferro privately intended.

It is also inaccurate to imagine him personally concealing the method for the entire twenty years between about 1515 and 1535: he died in 1526. The technique's restricted circulation outlasted its discoverer.

Public challenges could have occupational consequences, but losing one did not automatically cancel a university appointment under a universal rule. The period offered no single modern career track for mathematicians. Teaching, medicine, engineering, books, court connections, and competitive demonstrations overlapped.

Tartaglia's own life makes those overlaps visible. His nickname refers to a speech impediment he attributed to wounds suffered during the French sack of Brescia in 1512. His account of hiding with family and being struck by a soldier is another personal narrative to identify as such, not an independently recorded medical file.

His Nova scientia of 1537 concerned artillery. In its prefatory presentation, he describes an earlier reluctance to improve a destructive art and a later decision to publish in the face of an Ottoman threat. Mathematics, livelihood, defense, and moral justification meet in the same book.

Cardano moved among medicine, teaching, universities, and powerful patrons. Ferrari would turn public mathematical competition into professional capital. None fits neatly into the image of a researcher whose obvious obligation was immediate journal publication.

This context does not settle later accusations of broken promises. It explains why a secret rule could be worth negotiating over. Publishing distributed an advantage to rivals; withholding preserved something its possessor alone might use. Before judging the ensuing dispute, we need to know which claims survive from which participants.

2. A Promise in One Man's Book

Cardano sought Tartaglia's cubic rule in 1539. Surviving correspondence describes approaches through intermediaries and proposals that included publication with attribution or promises of secrecy. Tartaglia initially resisted, wishing to publish his own work.

An invitation involving a powerful Milanese connection helped draw him to Milan. In Tartaglia's later narrative, the decisive meeting occurred at Cardano's house on March 25. He transmitted the rule in verse. The surviving mnemonic concerns operations for solving types of cubics, not a complete geometrical demonstration of every case.

The famous oath has a specific documentary status. Its earliest surviving full wording is in Tartaglia's 1546 book, seven years after the alleged private exchange. We have no jointly signed 1539 document or neutral minutes fixing every word of the promise.

Tartaglia presents Cardano as swearing by sacred and gentlemanly obligations never to publish the discovery and even to encode it so that others could not read it after his death. That is evidence of what Tartaglia publicly alleged in 1546. It is not an independent recording of the meeting.

Cardano's Ars magna, published in 1545, acknowledges receiving the rule from Tartaglia and does not claim the original cubic discovery entirely for himself. It traces priority to del Ferro, recognizes Tartaglia's independent rediscovery, and emphasizes the subsequent work of proof, generalization, and organization. Ferrari receives credit for the quartic method.

Publication order matters. Ars magna appeared before Tartaglia printed the full oath account. It cannot be described as Cardano's reply to that already-published accusation. Ferrari's public interventions, beginning in 1547, belong to the subsequent controversy.

The traditional explanation for Cardano's decision to publish involves his and Ferrari's visit to Bologna around 1543, where they consulted material preserved through del Ferro's son-in-law, Annibale della Nave. The broad episode is generally accepted, but its minute details rely heavily on later contested reports. The complete original notebook does not survive for us to reopen.

Two ethical arguments can be distinguished without endorsing either as a decisive verdict. Cardano's defenders treat the discovery of an earlier source as removing Tartaglia's exclusive claim over the method and point out that suppressing the cubic also obstructed Ferrari's independent quartic work.

Tartaglia's position treats the promise as a personal obligation unaffected by someone else's earlier discovery. Prior discovery does not necessarily release the recipient of a confidential communication from his undertaking. The alleged release is an argument, not a surviving agreed clause.

Ferrari issued six public challenge sheets, and Tartaglia printed six replies during 1547–48. The exchange, including sixty-two problems, survives in a form that permits close comparison. The public quarrel is not a later novelist's invention.

Their eventual encounter in Milan is conventionally dated August 10, 1548. Standard accounts judge Ferrari to have emerged stronger and Tartaglia to have withdrawn. The detailed theater often supplied around it—crowds, official judges, entourages, and a furtive overnight departure—is much less securely documented.

What we know about the actual proceedings is limited and dependent on interested reports. A historian should not let the abundant printed preliminaries create the illusion of an equally abundant neutral record of the meeting.

Print altered the evidence. Accusations, dates, insults, and mathematical questions could be duplicated and preserved. A 1546 allegation could not simply become a document created at the 1539 encounter. A 1545 edition remained a witness to what had been published before the later attack.

Even printed dates are not literally unalterable: editions, reissues, variants, and calendar systems require bibliographical checking. The gain was durable, publicly comparable witnesses, not an infallible timestamp or a fully formed modern priority system.

The archive is consequently skewed in an intelligible way. The loud public dispute was designed to circulate. The private promise was not. We know much about the quarrel's published forms and much less about the precise commitment at its center. The best-preserved part of a controversy need not be the part that would settle it.

3. Thirteen Forms

Ars magna appeared at Nuremberg in 1545 in forty chapters. Chapters XI through XXIII treat thirteen forms of cubic equations, and the quartic material also distinguishes forms that modern notation can unify.

This is not pointless repetition. The classification reflects conventions about positive quantities and which side of a relation contains a term. Moving a term across an equality with a changed sign may be routine now; in that presentation it can move the problem into a different named type.

As with al-Khwarizmi's six forms, a modern polynomial written with arbitrary signed coefficients can conceal the historical work performed by the classification. Cardano knew and used negative quantities, but his system did not simply reproduce today's uniform coefficient convention.

A famous passage in Chapter XXXVII asks for two parts of ten whose product is forty. If the parts must be real, the request cannot be satisfied. Writing them as 5 + t and 5 − t makes the obstruction visible: their product is 25 − t², no greater than twenty-five for real t.

Cardano nevertheless continues the formal calculation. In modern notation the two expressions are 5 + √(−15) and 5 − √(−15). Their sum is ten, and formal multiplication gives 25 − (−15), or forty.

The important event is not that he already possesses the modern complex field. He is willing to manipulate expressions lacking the ordinary magnitude interpretation expected of an answer. He labels the work with language often translated as sophistic and comments on its extreme subtlety and limited usefulness.

The behavior exceeds the dismissive vocabulary. The expressions are retained long enough to satisfy the formal conditions. A provisional category gives an unfamiliar operation space to work, even while the author declines to treat its objects as ordinary numbers.

That is similar to the role of fictitious negative quantities in the previous essay, but the distinction must remain clear. A negative quantity and the square root of a negative quantity create different mathematical problems. Historical labels overlap without making the objects identical.

The example is quadratic. It is not evidence that Cardano successfully used nonreal intermediate quantities to resolve the difficult three-real-root cases in the cubic chapters. Popular accounts often transfer the success from one setting to another.

The correspondence published by Tartaglia indicates that Cardano encountered troubling cubic examples soon after receiving the rule in 1539. He questioned the operation; Tartaglia's reply attributed the difficulty to Cardano's handling and reflected growing suspicion about the exchange.

Again, the letters reach us through an interested later publication. Their mathematical content can be examined separately from accepting every implication of the editorial narrative.

Ars magna contains cubics that modern analysis places in the three-real-root regime. But where the radical procedure leads to square roots of negative quantities, Cardano does not consistently carry those expressions through to a successful resolution in the manner later associated with Bombelli.

Three things should therefore be separated. Cardano recognizes an obstruction in the rule. In a different problem he experiments with the formal arithmetic beyond the obstruction. Another author develops that arithmetic more systematically and brings it back into the cubic calculation.

The distinction matters because mathematical possibility rarely arrives in one indivisible moment. An operation can be tried before its general rules are organized; the rules can be organized before their objects receive an accepted interpretation; an interpretation can be published before it reaches a substantial audience.

The next section identifies the obstruction itself. Its force is stronger than the observation that an equation sometimes has an answer a mathematician does not like. Here the desired answers can all be real, while a particular successful-looking route insists on leaving the real domain along the way.

4. The Irreducible Case

Start with a cubic whose quadratic term has been removed by a change of variable. In modern notation write it as x³ + px + q = 0. Set x = u + v. Expanding gives u³ + v³ + (3uv + p)(u + v) + q = 0.

Choose 3uv + p = 0. Then the remaining requirements are u³ + v³ = −q and u³v³ = −(p/3)³. The two cubes are therefore the roots of a quadratic equation.

This modern symbolic derivation is an explanatory reconstruction, not a facsimile of Cardano's notation. It makes the difficulty inspectable: the auxiliary values are −q/2 ± √((q/2)² + (p/3)³).

When the cubic has one simple real root and a nonreal pair, the expression under that square root is positive. Real cube roots of the two real auxiliary values can produce the real solution. Boundary cases require their own treatment.

When the cubic has three distinct real roots, the same expression is negative. The auxiliary quadratic has nonreal conjugate solutions. Along this radical route, finding real roots requires first calculating with nonreal quantities.

Appropriate paired cube roots then recombine to give a real result. The pairing matters: independently choosing arbitrary cube roots need not satisfy uv = −p/3. The other coordinated choices generate the other real roots.

The nonreal pieces do not contaminate the answer. They encode relationships that cancel appropriately when the pieces are assembled. The destination lies on the real line, while this representation of the journey leaves it.

That is a powerful reason to study the unfamiliar operations. The difficulty is not solved by announcing that the original problem has no real answer. It has three. Nor can one merely change a sign in the formula and assume the obstruction disappears.

But the conclusion is about a route, not every possible method. Trigonometric expressions can represent these roots within real quantities. Numerical methods can approximate them. Geometrical constructions may locate them. Some special cubics can be factored immediately.

The familiar phrase casus irreducibilis has a stronger theorem behind it when irreducibility is specified. For a cubic with rational coefficients that is irreducible over the rationals and has three real roots, the roots cannot be expressed from the coefficients by real radicals alone in the usual radical framework. Wantzel's 1843 work established the classical obstruction.

Irreducibility is indispensable. If a rational root is already available, factorization reduces the remaining problem to a quadratic, and real expressions can be obtained without following Cardano's nonreal detour.

This also means that Bombelli's famous example, which has the known root four, illustrates the arithmetic of the detour but is not itself an irreducible cubic over the rationals. It should not be used as though it proved the later theorem's full hypothesis.

The historical pressure and the mathematical theorem belong to different dates. Sixteenth-century practitioners encountered the unwanted square roots in their formula. Nineteenth-century algebra could state and demonstrate why a specified class of alternative real-radical routes would not solve the general difficulty.

Calling the formula the only possible road in the sixteenth century would be too strong even historically. Methods of approximation and geometrical investigation existed. What made it consequential was its status as a new general algebraic procedure whose own internal requirements exposed the limitation.

The essay's theme is therefore deliberately precise. Sometimes an object lacking accepted status appears not as an optional fantasy but as a necessary intermediate within a valued method. The method's usefulness does not prove the object's entire theory. It creates a concrete problem that refusing to look at the intermediate expression does not resolve.

5. An Engineer's Arithmetic

Rafael Bombelli, baptized at Bologna in 1526 and dead in 1572, did not follow a conventional university mathematics career. His documented world included engineering, surveying, drainage, land boundaries, and patrons concerned with making terrain usable.

He trained with an architect and engineer. Records connect him with Tuscan land-reclamation work during the middle of the century, with interruptions and resumptions typical of ambitious projects. He also participated in a proposed bridge intervention at Rome and advised on drainage schemes.

Calling him merely an amateur mathematician can obscure how much specialized quantitative work belonged outside university chairs. Calling his algebra a consequence of marsh engineering would go beyond the evidence. The two activities belong to one career without a demonstrated psychological chain from one to the other.

His Algebra developed over years. He began writing during the 1550s, planned five books, and published the first three in 1572. The geometrical material of the remaining books stayed in manuscript, was rediscovered in 1923, and was edited in 1929.

Bombelli anticipated resistance. He described his own initial suspicion that the unfamiliar operations were closer to sophistry than truth and warned readers that the subject demanded concentrated attention. This is not the language of someone to whom the objects were already ordinary.

Yet he continued. He gave special names to opposed forms associated with square roots of negatives, conventionally rendered through phrases such as plus of minus and minus of minus. Those expressions do not amount to our symbol i, but they allow distinctions that ordinary verbal arithmetic did not handle comfortably.

He then gave operational rules. In modern terms, like imaginary units multiply to a negative real unit, while opposite ones multiply to a positive real unit. Real and nonreal parts can be combined systematically in addition and multiplication.

A multiplication table does not by itself provide a modern foundational construction or a global consistency proof. It does establish a disciplined calculational practice that can be tested across examples. That is already more than tolerating one anomalous radical and hoping for the best.

His celebrated example is x³ = 15x + 4. The value x = 4 plainly works. Cardano's expression gives ∛(2 + √(−121)) + ∛(2 − √(−121)), or, in modern notation, ∛(2 + 11i) + ∛(2 − 11i).

Bombelli's decisive move is to seek the complicated cube roots in a simpler paired form. Using the rules, (2 + i)³ = 2 + 11i and (2 − i)³ = 2 − 11i. Their sum is four.

The calculation can be checked directly. Squaring 2 + i gives 3 + 4i; multiplying again by 2 + i gives 6 + 11i + 4i² = 2 + 11i. The conjugate calculation reverses the nonreal part. The unfamiliar terms disappear only after doing indispensable work within the expression.

The answer's novelty is not the point: four was already known. The value of the example is that the proposed rules carry the calculation through the obstruction and return the expected real solution without an arbitrary change of procedure midway.

Bombelli did not thereby prove that every conceivable manipulation with these objects was safe. Branch choices and extension of identities would continue to cause difficulties. He demonstrated a coherent and productive body of operations, not a modern theorem about the consistency of an entire field.

Nor did the publication instantly persuade every European mathematician. We lack a dependable first-edition sales total and a complete map of early readership. Later praise and influence cannot establish a large audience in the 1570s.

Some historical accounts emphasize Bombelli's conceptual importance; others caution that his immediate circulation was narrower than his later canonical position suggests. Both points can be true. An achievement can be mathematically substantial before it becomes widely known.

What changed on the page was enough to matter. A square root of a negative quantity no longer served only as a warning to stop. It could enter a rule-governed language that successfully mediated between real conditions and a real answer. Explaining what that language's objects were would take further work.

6. Two and a Half Centuries

From Bombelli to the early nineteenth century, the history is not well summarized as universal rejection. Nonreal quantities became increasingly useful in equations, logarithms, trigonometry, series, and analysis. Euler's 1748 connection between exponential and trigonometric functions is a prominent landmark.

Operational fluency and ontological reassurance did not always move together. A mathematician might call a quantity impossible in one discussion and use it skillfully in another. The words can mark a problem of interpretation without signaling a ban on calculation.

Descartes's Geometry of 1637 helped fix the term imaginary. Discussing roots, he distinguishes what can be counted formally as possible roots from quantities actually corresponding to them. Imaginary indicates the lack of an accepted real-magnitude counterpart, not a romantic tribute to creativity.

The term should not be reduced to a personal insult either. Its force comes from a conceptual contrast between real magnitude and something only posited in the algebraic accounting. Later readers inherit the ordinary-language implication after the mathematical framework has changed.

Wallis attempted geometrical interpretations in the seventeenth century, including perpendicular directions. These were important steps, but a suggestive perpendicular construction is not yet a fully coordinated representation of every complex number, addition, multiplication, length, and angle.

Leibniz's discussions likewise combine extensive use with language of fictions and uncertain modes of being. Such expressions should be read with their mathematical purposes rather than gathered as proof that nothing was understood before a plane was drawn.

Euler's use of i requires another date distinction. The notation is associated with a paper written in 1777 but published in 1794. Composition and public availability are separate historical events.

Euler's proficiency also did not eliminate errors caused by carrying familiar identities beyond their valid domains. Square-root multiplication, harmless under suitable nonnegative real hypotheses, becomes branch-sensitive over complex numbers. Earlier caution and later expertise do not form a simple ladder on which every successor is uniformly safer.

Caspar Wessel, a surveyor and cartographer, supplied a systematic geometrical account in work submitted to the Danish academy in 1797 and published in 1799. Directed segments represent the quantities; addition follows composition of directed lengths, while multiplication multiplies lengths and adds angles.

The representation was a major achievement, yet its international reception was limited for decades. Language, publication venue, and scholarly networks mattered. Renewed historical attention in the late nineteenth century, including a French translation in 1897, restored recognition of Wessel's priority.

Jean-Robert Argand independently developed a related account in work associated with 1806. The exact circumstances of its initial private circulation are less secure than later textbook narratives imply. Discussion in Gergonne's journal around 1813 brought the approach into a wider French mathematical network.

Gauss's relation to the subject spans several stages. His 1799 dissertation on the fundamental theorem of algebra belongs to the background, while his especially influential public discussion of complex arithmetic appeared in the early 1830s. The familiar date 1831 concerns the announcement and exposition associated with his work on biquadratic residues; publication histories should not be flattened into one event.

Hamilton pursued a complementary route in the 1830s. Instead of asking where a mysterious square root lived, he organized complex quantities as pairs of real numbers. Addition acts componentwise. Multiplication is defined so that (a,b)(c,d) = (ac − bd, ad + bc).

Under those rules, (0,1) multiplied by itself is (−1,0). The imaginary unit need not be assumed beforehand as an inexplicable root; it becomes a specific element in a constructed system. Hamilton's work developed around 1833 and appeared in mature published form in 1837.

A geometrical picture helps one see the operations. A pair construction shows how to build them from accepted real quantities and verify their laws. Neither makes two centuries of earlier analysis irrelevant; each answers questions that previous practice had kept open.

The timeline defeats the single-cause story in which nobody believed until someone drew the complex plane. Extensive successful use preceded the mature representation. Wessel's limited early reception shows that publishing a representation did not automatically produce acceptance.

Geometry was both a product of accumulated practice and a contributor to further legitimacy. Formal constructions, applications, notation, and circulation all mattered. The gray interval contained partial explanations and real conceptual work, not merely two hundred years of blind calculation.

7. Direct, Inverse, Lateral

In his early-1830s discussion, Gauss complained that names such as positive, negative, imaginary, and impossible helped create unnecessary obscurity. A vocabulary of direct, inverse, and lateral units, he suggested, might have made the relations less mysterious.

The remark identifies naming as part of the problem. A word can continue to imply illegitimacy after a stable mathematical framework has made its objects entirely ordinary. The inherited contrast between real and imaginary is unusually effective at misleading newcomers.

The labels nevertheless need chronological and conceptual separation. Cardano's sophistic language belongs to the 1545 experiment. Bombelli uses related language when describing his initial doubts. Descartes's imaginary becomes the durable technical term. Impossible occurs in later discussions with varying force.

Absurd belongs especially to the negative-number history examined in the previous essay. It should not be assigned indiscriminately to mature pure imaginary numbers or projected into a fifteenth-century discussion that concerned something else.

This distinction does not make the histories unrelated. If negative quantities themselves lacked an uncontested interpretation in a framework of ordinary magnitude, their square roots would be still harder to place. But the second difficulty is not merely a stronger emotional reaction to the first.

Why did imaginary remain? We do not have a mathematical congress voting to preserve a bad name. Terminological inertia is a plausible explanation: real and imaginary parts, axes, and quantities became embedded in teaching and research across languages.

That explanation is an inference about linguistic stability, not a documented universal decision. Technical terms also routinely shed part of their everyday meaning. A mathematician using imaginary today does not thereby deny the existence or legitimacy of a complex number.

Changing a term has coordination costs. Textbooks, translations, paired expressions, and reference works have to change together. Removing one misleading implication may not outweigh the practical advantage of a familiar shared vocabulary. That is a reasonable account of persistence, not a proof that nobody ever proposed alternatives.

Chinese terminology requires a separate investigation. Modern xushu and shishu form an imaginary/real pair and share written forms with Japanese technical vocabulary. Shared characters do not establish which language coined a term first or the precise direction of every transfer.

Nineteenth-century Chinese algebra translations preceded some of the later mass borrowing of Japanese technical vocabulary. Sino-Japanese intellectual exchange was not one-way. A responsible first-use claim would require comparison of dated editions, not a guess based on how the compound looks.

This review does not establish a secure first Chinese coiner, first edition, or complete series of competing translations. The absence should remain explicit. It is better to leave a bibliographical question open than to populate it with a plausible famous translator.

The semantic observation can be made more cautiously. Xu carries associations of emptiness or nonactuality that can echo the contrast encoded by imaginary. That may help explain why the term feels misleading to modern learners. It is not evidence of the translator's private intention or of a universal rule that only accurate translations survive.

The previous essay followed old Chinese words for positive and negative into new algebraic settings. Here the exact naming route is less fully established. The asymmetry is part of the evidence, not a defect to conceal by giving both stories matching origin dates.

Gauss's complaint remains pertinent without resolving the etymology of every language. The conceptual system can change faster than its ordinary-language labels. A student may receive an obsolete philosophical hint before learning the modern definition that makes the hint unnecessary.

8. Going Through

The negative quantities of the previous essay often arrived with ready interpretations: debt, loss, or reversed direction. Square roots of negatives presented a different pressure. In important algebraic calculations, intermediate expressions appeared for which no accepted ordinary-magnitude counterpart was available.

The cubic case makes the contrast vivid. The desired roots can all be real. The particular radical representation nonetheless passes through nonreal quantities before returning a real value. The intermediate objects do not have to remain visibly in the final answer to have performed mathematical work.

This is not their only historical role. Cardano's sum-and-product example gives nonreal final solutions in modern terms, and later analysis treats complex quantities as objects of direct investigation rather than merely temporary scaffolding. The detour is an important entry route, not an exhaustive definition.

Nor was everyone compelled to use it or everyone united in refusing its legitimacy. Approximation, geometry, and special factorizations supplied alternatives in various problems. The argument concerns a specific pressure within general algebraic methods, followed by a widening body of uses.

The source's strongest rhetorical version—that for two hundred and fifty years nobody offered any worthwhile explanation—would erase Wallis and the other interpretive efforts already described. What remained absent for much of the interval was a broadly accepted, systematically coordinated account, not every meaningful idea.

Bombelli's rules supplied a disciplined operational foothold. They did not establish a modern consistency theorem. Later geometrical accounts connected the operations with lengths and angles. Pair constructions showed how a system could be explicitly made from real components. These achievements answered different questions.

The distribution of credit therefore should not be collapsed into one inventor. Del Ferro, Tartaglia, Cardano, and Ferrari belong to the history of the equations and their disputed circulation. Bombelli develops the arithmetic at a crucial obstruction. Wessel, Argand, Gauss, Hamilton, and others contribute interpretations, constructions, and dissemination.

Even this list is selective. Its purpose is not to replace one heroic origin story with a slightly longer roll of heroes. It is to show why discovery, demonstration, interpretation, publication, and acceptance have different evidence and dates.

The social settings are equally varied. A university teacher guards a rule; an independent instructor publishes a controversy; a physician organizes a major algebra; an engineer develops nonreal arithmetic; a surveyor supplies a geometrical representation that receives little international attention for decades.

Those occupations are not colorful distractions from the mathematics. They help explain what could be communicated, what was worth keeping secret, and which readers were likely to encounter a publication. But they do not license speculative claims that a marsh engineer's practical worldview caused his abstract innovation.

The philosophical interpretation offered by this series is that successful work can make a conceptual boundary increasingly expensive to maintain. If a useful procedure repeatedly encounters an excluded object, mathematicians may revise the procedure, restrict its claims, or enlarge the domain in which it operates.

In this case, enlargement became exceptionally fruitful. The unfamiliar expressions did not need to be disguised as ordinary positive magnitudes. They could acquire rules, representations, and explicit constructions that explained why the successful operations belonged together.

A pair such as (0,1) is no less definite because it corresponds to what earlier writers called imaginary. A point in a plane is no less available to mathematical reasoning because a historical name still contrasts it with real. The modern account does not vindicate every earlier informal manipulation; it makes their successes and failures easier to distinguish.

The name nevertheless continues to do pedagogical damage. A beginner hears one kind of number called real and another imaginary and may infer that one exists while the other is a useful pretense. The seventeenth-century contrast concerned an accepted kind of magnitude. Modern mathematical existence is not restricted to that category.

Three centuries of conceptual work separate those meanings. The persistence of the word conceals the change until someone explains it. That is why naming belongs to this history, though it cannot replace the history of the operations.

The lesson is not simply that mathematics may imagine anything it likes. New objects require coherent definitions and justified laws. The more interesting claim is that usefulness can disclose where old definitions are too restrictive—and that proving a new system's legitimacy need not consist in forcing it back into the old interpretation.