I have recently read the Wikipedia article on Abuse of Notation (this link is to the version of 29 December 2011, since I will eventually edit it). The Handbook of Mathematical Discourse and abstractmath.org mention this idea briefly. It is time to expand the abstractmath article and to redo parts of the Wikipedia article, which contains some confusions.

This is a preliminary draft, part of which I’ll incorporate into abstractmath after you readers make insightful comments :).

The phrase “Abuse of Notation” is used in articles and books written by research mathematicians. It is part of Mathematical English. This post is about

- What “abuse of notation” means in mathematical writing and conversation.
- What it
*could*be used to mean. - Mathematical usage in general. I will discuss this point in the context of the particular phrase “abuse of notation”, not a bad way to talk about a subject.

## Mathematical Usage

### Sources

If I’m going to write about the *usage* of Mathematical English, I should ideally verify what I claim about the usage by finding **citations** for a claim: *documented quotations that illustrate the usage*. This is the standard way to produce any dictionary.

There is no complete authoritative source for usage of words and phrases in Mathematical English (ME), or for that matter for usage in the Symbolic Language (SL).

- The
[2] covers technical terms and symbols used in school math and in much of undergraduate math, but not so much of research math. It does not mention being based on citations and it hardly talks about usage at all, even for notorious student-confusing notations such as “$latex \sin^k x$”. But it appears quite accurate with good explanations of the math it covers.*Oxford Concise Dictionary of Mathematics* - I wrote
**Handbook of Mathematical Discourse**to stimulate investigations into mathematical usage. It describes a good many usages in Mathematical English and the Symbolic Language, documented with citations of quotations, but is quite incomplete (as I said in its Introduction). The Handbook has 428 citations for various usages. (They are at the end of the on-line PDF version. They are not in the printed book, but are on the web with links to pages in the printed book.) **MathWorld**has an extensive list of mathematical words, phrases and symbols, and accurate definitions or descriptions of them, even for a great many advanced research topics. It also frequently mentions usage (see formula and inverse sine), but does not give citations.**Wikipedia**has the most complete set of definitions of mathematical objects that I know of. The entries sometimes mention usage. I have not detected any entry that gives citations for usage. Not that that should stop anyone from adding them.

### Teaching mathematical usage

In explaining mathematical usage to students, particularly college-level or higher math students, you have choices:

- Tell them what you
*think*the usage of a word, phrase, or symbol is, without researching citations. - Tell them what you think the usage
*ought*to be. - Tell them what you think the usage is, supported by citations.

(1) has the problem that you can be wrong. In fact when I worked on the Handbook I was amazed at how wrong I could be in what the usage was, in spite of the fact that I had been thinking about usage in ME and SL since I first started teaching (and kept a folder of what I had noticed about various usages). However, professional mathematicians generally have a reasonably accurate idea about usage for most things, particularly in their field and in undergraduate courses.

(2) is dangerous. Far too many mathematicians (but nevertheless a minority), introduce usage in articles and lecturing that is* not common *or that they invented themselves. As a result their students will be confused in trying to read other sources and may argue with other teachers about what is “correct”. It is a *gross violation of teaching ethics* to tell the students that (for example) “*x* > 0″ allows *x* = 0 and not mention to them that nearly all written mathematics does *not *allow that. (Did you know that a small percentage of mathematicians and educators *do *use that meaning, including in some secondary institutions in some countries? It is partly Bourbaki’s fault.)

(3) You often *can’t* tell them what the usage is, supported by citations, because, as mentioned above, documented mathematical usage is sparse.

I think people should usually choose (1) instead of (2). If they do want to introduce a new usage or notation because it is “more logical” or because “my thesis advisor used it” or something, they should *reconsider*. Most such attempts have failed, and thousands of students have been confused by the attempts.

## Abuse of notation

“Abuse of notation” is a phrase used in mathematical writing to describe terminology and notation that does not have **transparent meaning**. (Transparent meaning is described in some detail under “compositional” in the Handbook.)

Abuse of notation was originally defined in French, where the word “abus” does not carry the same strongly negative connotation that it does in English.

### Suppression of parameters

One widely noticed practice called “abuse of notation” is the use of the name of the underlying set of a mathematical structure to refer to a structure. For example, a group is a structure $latex (G,\text{*})$ where $latex G$ is a set and * is a binary operation with certain properties. The most common way to refer to this structure is simply to call it $latex G$. Since any set of cardinality greater than 1 has more than one group structure on it, this does not include all the information needed to determine the group. This type of usage is cited in **82** below. It is an example of suppression of parameters.

Writing “$latex \log x$” without mentioning the base of the logarithm is also an example of suppression of parameters. I think most mathematicians would regard this as a convention rather than as an abuse of notation. But I have no citations for this (although they would probably be easy to find). I doubt that it is possible to find a rational distinction between “abuse of notation” and “convention”; it is all a matter of what people are used to saying.

### Synecdoche

The naming of a structure by using the name of its underlying set is also an example of synecdoche, the naming of a whole by a part (for example, “wheels” to mean a car).

Another type of synecdoche that has been called abuse of notation is referring to an equivalence class by naming one of its elements. I do not have a good quotation-citation that shows this use. Sometimes people write 2 + 4 = 1 when they are working in the Galois field with 5 elements. But that can be interpreted in more than one way. If GF[5] consists of equivalence classes of integers (mod 5) then they are indeed using 2 (for example) to stand for the equivalence class of 2. But they could instead define GF[5] in the obvious way with underlying set {0,1,2,3,4}. In any case, making distinctions of that sort is pedantic, since the two structures are related by a natural isomorphism (next paragraph!)

### Identifying objects via isomorphism

This is quite commonly called “abuse of notation” and is exemplified in citations **209, 395 **and** AB3.
**

### Overloaded notation

John Harrison, in [1], uses “abuse of notation” to describe the use of a function symbol to apply to both an element of its domain and a subset of the domain. This is an example of overloaded notation. I have not found another citation for this usage other than Harrison and I don’t remember anyone using it. Another example of overloaded notation is the use of the same symbol “$latex \times$” for multiplication of numbers, matrices and 3-vectors. I have never heard that called abuse of notation. But I have no authority to say anything about this usage because I haven’t made the requisite thorough search of the literature.

### Powers of functions

The Wikipedia Article on abuse of notation (29 Dec 2011 version) mentions the fact that $latex f^2(x)$ can mean either $latex f(x)f(x)$ or $latex f(f(x))$. I have never heard this called abuse of notation and I don’t think it should be called that. The notation “$latex f^2(x)$” can in ordinary usage mean one of two things and the author or teacher should say which one they mean. *Many* math phrases or symbolic expressions can mean more than one thing and the author generally should say which. I don’t see the point of calling this phenomenon abuse of notation.

### Radial concept

The Wikipedia article mentions phrases such as “partial function”. This article* does* provide a citation for Bourbaki for calling a sentence such as “Let $latex f:A\to B$ be a partial function” abuse of notation. Bourbaki is wrong in a deep sense (as the article implies). There are several points to make about this:

- Some authors, particularly in logic,
*define*a function to be what most of us call a partial function. Some authors require a ring to have a unit and others don’t. So what? - The phrase “partial function” has a standard meaning in math: Roughly “it is a function except it is defined on only part of its domain”. Precisely, $latex f:A\to B$ is a partial function if it is a function $latex f:A’\to B$ for some subset $latex A’$ of $latex A$.
- A partial function is not in general a function. A stepmother is not a mother. A left identity may not be an identity, but the phrase “left identity” is defined precisely. An incomplete proof is not a proof, but you know what the phrase means! (Compare “expectant mother”).
*This is the way we normally talk and think*. See the article “radial concept” in the Handbook.

### Other uses

**AB4** involves a redefinition of “$latex \in$” in a special case. Authors redefine symbols all the time. This kind of redefinition on the fly probably should be avoided, but since they did it I am glad they mentioned it.

I have not talked about some of the uses mentioned in the Wikipedia article because I don’t yet understand them well enough. **AB1** and **AB2** refer to a common use with pullback that I am not sure I understand (in terms of how they author is thinking of it). I also don’t understand **AB5**. Suggestions from readers would be appreciated.

## Kill it!

Well, it’s more polite to say, we don’t need the phrase “abuse of notation” and it should be deprecated.

- The use of the word “abuse” makes it sound like a bad thing, and most instances of abuse of notation are nothing of the sort. They make mathematical writing much more readable.
- Nearly everywhere it is used it could just as well be called a convention. (This requires verification by studying math texts.)

## Citations

The first three citations at in the Handbook list; the numbers refer to that list’s numbering. The others I searched out for the purpose of this post.

**82.** Busenberg, S., D. C. Fisher, and M. Martelli (1989), *Minimal periods of discrete and smooth orbits*. American Mathematical Monthly, volume 96, pages 5–17. [p. 8. Lines 2–4.]

Therefore, a normed linear space is really a pair $latex (\mathbf{E},\|\cdot\|)$ where $latex \mathbf{E}$ is a linear vector space and $latex \|\cdot\|:\mathbf{E}\to(0,\infty)$ is a norm. In speaking of normed spaces, we will frequently abuse this notation and write $latex \mathbf{E}$ instead of the pair $latex (\mathbf{E},\|\cdot\|)$.

**209.** Hunter, T. J. (1996), *On the homology spectral sequence for topological Hochschild homology*. Transactions of the American Mathematical Society, volume 348, pages 3941–3953. [p. 3934. Lines 8–6 from bottom.]

We will often abuse notation by omitting mention of the natural isomorphisms making $latex \wedge$ associative and unital.

**395**. Teitelbaum, J. T. (1991), ‘The Poisson kernel for Drinfeld modular curves’. Journal of the American Mathematical Society, volume 4, pages 491–511. [p. 494. Lines 1–4.]

$latex \ldots$ may find a homeomorphism $latex x:E\to \mathbb{P}^1_k$ such that $latex \displaystyle x(\gamma u) = \frac{ax(u)+b}{cx(u)+d}$. We will tend to abuse notation and identify $latex E$ with $latex \mathbb{P}^1_k$ by means of the function $latex x$.

**AB1**. Fujita, T. *On the structure of polarized manifolds with total deficiency one*. I. J. Math. Soc. Japan, Vol. 32, No. 4, 1980.

Here we show examples of symbols used in this paper $latex \ldots$

$latex L_{T}$: The pull back of $latex L$ to a space $latex T$ by a given morphism $latex T\rightarrow S$ . However, when there is no danger of confusion, we OFTEN write $latex L$ instead of $latex L_T$ by abuse of notation.

**AB2**. Sternberg, S. *Minimal coupling and the symplectic mechanics of a classical*

* particle in the presence of a Yang-Mills field.* Physics, Vol. 74, No. 12, pp. 5253-5254, December 1977.

On the other hand, let us, by abuse of notation, continue to write $latex \Omega$ for the pullback of $latex \Omega$ from $latex F$ to $latex P \times F$ by projection onto the second factor. Thus, we can write $latex \xi_Q\rfloor\Omega = \xi_F\rfloor\Omega$ and $latex \ldots$

**AB3.** Dobson, D, and Vogel, C. *Convergence of an iterative method for total variation denoising.* SIAM J. Numer. Anal., Vol. 34, pp. 1779, October, 1997.

Consider the approximation

(3.7) $latex u\approx U\stackrel{\text{def}}{=}\sum_{j=1}^N U_j\phi_j$ $latex \ldots$

In an abuse of notation, $latex U$ will represent both the coefficient vector $latex \{U_j\}_{j=1}^N$ and the corresponding linear combination (3.7).

**AB4.** Lewis, R, and Torczon, V. *Pattern search algorithms for bound constrained minimization. * NASA Contractor Report 198306; ICASE Report No. 96-20.

By abuse of notation, if $latex A$ is a matrix, $latex y\in A$ means that the vector $latex y$ is a column of $latex A$.

**AB5.** Allemandi, G, Borowiecz, A. and Francaviglia, M. *Accelerated Cosmological Models in Ricci squared Gravity*. ArXiv:hep-th/0407090v2, 2008.

This allows to reinterpret both $latex f(S)$ and $latex f'(S)$ as functions of $latex \tau$ in the expressions:

$latex \begin{equation*}\begin{cases}

f(S) = f(F(\tau)) = f(\tau )\\

f'(S) = f'(F(\tau )) = f'(\tau )\end{cases}\end{equation*}$

following the abuse of notation $latex f(F(t )) = f(t )$ and $latex f'(F(t )) = f'(t )$.

## References

[1] Harrison, J. *Criticism and reconstruction, *in* Formalized Mathematics *(1996)*.*

[2] Clapham, C. and J. Nicholson. Oxford Concise Dictionary of Mathematics, Fourth Edition (2009). Oxford University Press.