Math's Alphabetical Author Order: Origins

Mathematics is the field most consistently associated with alphabetical-by-surname author order. Bibliometric analysis has found alphabetical ordering in roughly 94% of multi-author mathematics papers once the baseline chance of a coincidentally alphabetical list is factored out — with the rate running even higher on two-author papers. This guide covers where that norm actually came from, why it persisted for most of the twentieth century, and the specific, documented ways it is now under pressure — from citation-fairness research, from contribution-disclosure norms borrowed from other fields, and from large collaborative subfields where a strict alphabetical list has started to feel impractical.

For how mathematics’ convention compares to the position-coded system used in biomedicine or the randomization tool economics built in response to the same fairness problem, see How Authorship Order Is Decided: Conventions Across Disciplines. This page goes deeper on mathematics specifically: its origin story, the mechanics of why it took hold, and the evidence for how it is changing.

The origin: the “Hardy–Littlewood Rule”

The alphabetical convention in mathematics is widely traced to the working partnership between G. H. Hardy and J. E. Littlewood, two Cambridge number theorists and analysts whose joint output over several decades in the early twentieth century made them one of the most productive collaborations in the field’s history. Mathematical folklore — recounted in later writing about the pair, including a widely cited set of “rules” attributed to their working relationship — holds that they agreed on ground rules for collaborating, one of which was that joint papers would always carry both names regardless of who had actually done more of the work on any given paper, with the names listed alphabetically rather than ranked. That convention is still informally called the Hardy–Littlewood Rule in discussions of mathematical authorship.

The mechanism behind why this took hold field-wide, not just between two collaborators, is straightforward: once alphabetical order is understood field-wide to carry no information about relative contribution, using any other order becomes a signal in itself — putting a specific name first would imply that person did more, which most mathematical collaborators explicitly wanted to avoid asserting. Mathematics also differs structurally from lab-based sciences in ways that reinforced the norm: proofs are frequently the product of long, iterative, hard-to-apportion joint reasoning rather than separable tasks (running an experiment, writing code, drafting text) that a first/last-author hierarchy is built to signal. Where contribution genuinely cannot be cleanly separated, ranking by it is both difficult to do fairly and, by the field’s own logic, not obviously meaningful to do at all.

Economics adopted essentially the same convention independently, for a similar reason — and some large-collaboration corners of physics and astronomy also list contributors alphabetically, though for a different, scale-driven reason (see below). None of these are the same historical lineage as the Hardy–Littlewood story specifically; mathematics is the field where that particular origin account is most directly told.

What the norm actually optimizes for — and what it costs

The stated rationale for alphabetical order is that it treats co-authorship as intellectually joint work: every listed author’s name carries equal weight, and no one has to negotiate, or be seen to negotiate, over billing. That is a genuine advantage over position-coded conventions, where disputes over who is “first author” are one of the most common sources of authorship conflict in fields that use them (see first author and last author).

The documented cost is a name-position effect that has nothing to do with contribution: research on citation practices — most prominently in economics, which shares the alphabetical convention with mathematics — has found that authors whose surnames fall early in the alphabet accrue a measurable citation and visibility advantage, because they are listed first on the byline and are the ones readers see in shorthand citation forms such as “Smith et al.” This effect compounds over a career: early recognition, invitations, and even tenure-committee visibility can trace partly back to where a name falls in the alphabet rather than to the work itself. Because mathematics and economics share the same ordering convention, most of the empirical citation-fairness literature on this specific effect comes out of economics, but the underlying mechanism — alphabetical position as an accidental proxy for prominence — applies identically wherever strict alphabetical order is the norm, including mathematics.

How it’s changing

A documented decline since around 2010

Alphabetical ordering has been gradually declining across several fields that traditionally used it as the default — including mathematics, economics, education, and philosophy — roughly since 2010, as more researchers in these fields move toward orderings that are meant to signal relative contribution instead, or add an explicit contribution statement alongside an author list that no longer does that signaling work by position alone. This is a real, corroborated bibliometric trend, not a prediction; it does not mean alphabetical order has stopped being the majority practice in mathematics — it remains dominant — but the share of papers using it has been moving in one direction.

Explicit contribution statements alongside author order

A second, related shift is the spread of author-contribution disclosure as something separate from author order itself — the same separation of concerns behind the CRediT taxonomy (ANSI/NISO Z39.104-2022) in the sciences more broadly. Where a paper’s byline stays alphabetical but is paired with a short statement of who did what, the field gets to keep the fairness benefit of not ranking names while still giving readers, hiring committees, and funders the contribution information that a strict alphabetical order cannot convey on its own. CRediT itself was developed for the sciences generally and is not mathematics-specific, and mathematics journals have been slower than biomedical journals to formally require it — but the underlying idea, that authorship order and contribution disclosure are two separate questions that don’t have to be solved by the same mechanism, is increasingly visible in how mathematics papers are written even where a formal taxonomy isn’t mandated by the journal.

Randomization as an explicit alternative

Economics — mathematics’ closest peer in sharing the alphabetical default — built a specific, adoptable fix worth knowing about even outside economics: the American Economic Association’s Author Randomization Tool lets co-authors generate a certified random order and mark the byline with a circled “R” symbol, publicly signalling that the order was randomized rather than either alphabetical or contribution-ranked. The proposal behind it — Debraj Ray and Arthur Robson’s 2018 American Economic Review paper “Certified Random: A New Order for Co-Authorship” — directly targets the alphabetical name-position bias described above by decoupling name order from both contribution and the alphabet. Mathematics has no equivalent field-wide mechanism as of this writing, but the tool illustrates a workable alternative that a field sharing the same underlying problem has already built and normalized.

Large collaborations: alphabetical order for a different reason

It is worth distinguishing the classical Hardy–Littlewood rationale from a second, unrelated reason some very large collaborative projects — particularly in high-energy physics and astronomy, fields adjacent to mathematics in using large alphabetical bylines — also list contributors alphabetically. When an author list runs into the hundreds or thousands, as on major particle-physics collaboration papers, ranking individuals by contribution is not attempted at all; authorship is instead assigned collaboration-wide under the group’s own membership rules in effect at submission, and names are typically listed alphabetically or by a collaboration-defined convention simply because no other ordering scheme is practical at that scale. This is a scale-driven convenience, not the equal-intellectual-contribution rationale mathematics uses for a handful of co-authors — the two should not be read as the same justification even though the resulting byline looks similar.

What this means for a mathematics co-authorship team today

Alphabetical order remains the safe, expected default for a mathematics paper with a small number of co-authors, and departing from it without explanation can itself read as a signal you may not intend to send.
If contribution needs to be conveyed — for hiring, tenure, or grant reporting, where a committee outside the field may not know that alphabetical order carries no ranking information in mathematics — a short explicit contribution statement alongside the byline solves that without abandoning alphabetical order itself. See co-first authorship for how adjacent fields handle a related signaling problem.
Agree on the convention explicitly and early, especially on an interdisciplinary team where a co-author from biomedicine or another position-coded field may otherwise assume alphabetical placement is a ranking. This is the same advice CASRAI gives for author order generally — see authorship order conventions across disciplines.

Frequently asked questions

Is alphabetical author order still the majority practice in mathematics?

Yes. It remains the dominant convention in mathematics by a wide margin — bibliometric studies consistently find it on the large majority of multi-author papers. What has changed is the trend line, not the current default: adoption has been gradually declining since roughly 2010, alongside growing use of explicit contribution statements, not a wholesale switch to a different ordering system.

Does alphabetical order mean all co-authors contributed equally?

Not necessarily, and mathematics does not claim it does in a strict sense — the convention signals that contribution is not being ranked by position, which is a different claim than asserting contributions were literally identical. Readers outside mathematics sometimes misread an alphabetical byline as a ranking (assuming the first-listed name did the most work, following the convention their own field uses), which is exactly the kind of cross-disciplinary confusion an explicit note can prevent.

Where does the term “Hardy–Littlewood Rule” come from?

It refers informally to the working arrangement between mathematicians G. H. Hardy and J. E. Littlewood, whose prolific early-twentieth-century Cambridge collaboration is credited in mathematical folklore with popularizing the idea that joint papers should carry both names, unranked, regardless of who contributed what on a given paper. It is not a formally adopted rule of any journal or society — it is a description of a norm that grew out of a specific, celebrated collaboration and then generalized across the field.

Is there a mathematics equivalent of the AEA’s randomization tool?

Not a field-wide one as of this writing. The American Economic Association’s certified randomization tool was built specifically for economics, which shares mathematics’ alphabetical default and the same citation-fairness concern behind it. Nothing prevents a mathematics co-authorship team from adopting an equivalent randomize-and-disclose approach informally, but there is no equivalent formal, field-endorsed mechanism in mathematics itself.

How does CRediT apply to mathematics papers?

CRediT (ANSI/NISO Z39.104-2022) is discipline-agnostic and can in principle be applied to a mathematics paper the same way it is applied elsewhere, but mathematics journals have generally been slower to formally require it than biomedical and life-science journals have. Where it is used in mathematics, it typically supplements rather than replaces alphabetical order — the byline stays alphabetical, and a separate CRediT statement carries the contribution information.