WHAT'S WRONG WITH THESE EQUATIONS?

N. David Mermin

A major impediment to writing phys- saulted by so many egregious viola- not. Beyond that, should fortune
ics gracefully comes from the need to tions of even these simple precepts smile upon you and others actually
embed in the prose many large pieces that I offer them in the hope that a have occasion to mention your analy-
of raw mathematics. Nothing in few sinners—not only writers, but sis in papers of their own, they will
freshman composition courses pre- copy editors, publishers of journals not think the better of you for forcing
pares us for the literary problems and even the authors of the math- them into such locutions as "the
raised by the use of displayed equa- ematics subsections of literary style second equation after (3.21)" or "the
tions. Our knowledge is acquired manuals—may read them and repent third unnumbered equation from the
implicitly by reading textbooks and the error of their ways, or even be top in the left-hand column on p.
articles, most of whose authors have inspired to further beneficial studies 2485." Even should you solipsistical-
also given the problem no thought. of the sadly neglected field of mathe- ly choose to publish in a journal both
When I was a graduate teaching matico-grammatics. unrefereed and unread, you might
assistant in a physics course for non- Rule 1 (Fisher's rule). This rule, subsequently desire (just for the rec-
scientists, I was struck by the excep- named after the savant who repri- ord) to publish an erratum, the grace-
tional clumsiness with which ex- manded me for abusing it when I ful flow of which could only be en-
tremely literate students who lacked was young and foolish, simply en- sured if you had adhered to Fisher's
the exposure even to such dubious joins one to number all displayed rule in your original manuscript.
examples treated mathematics in equations. The most common viola- Rule 2 (Good Samaritan rule). A
their term papers. The equations tion of Fisher's rule is the misguided Good Samaritan is compassionate and
stood out like droppings on a well- practice of numbering only those helpful to one in distress, and there is
manicured lawn. They were invaria- displayed equations to which the text nothing more distressing than having
bly introduced by the word "equa- subsequently refers back. I call this to hunt your way back in a manu-
tion," as in "Pondering the problem of heresy Occam's rule. Back in the script in search of Eq. (2.47) not
motion, Newton came to the realiza- days of pens, pencils and typewrit- because your subsequent progress re-
tion that the key lay in the equation
ers, use of Occam's rule was kept quires you to inspect it in detail, but
under control by the pain of having merely to find out what it is about so
F=ma." (1) to renumber everything by hand you may know the principles that go
To these innocents equations were whenever it was deemed wise to into the construction of Eq. (7.38).
objects, gingerly to be pointed at or add a reference to a hitherto-unre- The Good Samaritan rule says: When
poked, not inseparably integrated marked-upon equation. One often referring to an equation identify it by
into the surrounding prose. encountered papers displaying the a phrase as well as a number. No
Clearly people are not born know- results of the ungainly Fisherian- compassionate and helpful person
ing how to write mathematics. The Occamite compromise: Number all would herald the arrival of Eq. (7.38)
implicit tradition that has taught us displayed equations that you think by saying "inserting (2.47) and (3.51)
what we do know contains both good you might want to refer to. Now into (5.13)..." when it is possible to
strands and bad. One of my defects of that automatic equation numbering say "inserting the form (2.47) of the
character being a preference for form macros can act upon symbolic electric field E and the Lindhard form
over substance, I have worried about names, the barrier to full Occamism (3.51) of the dielectric function e into
this over the years, collecting princi- has been removed, and it is neces- the constitutive equation (5.13)
ples that ought to govern the mar- sary to state emphatically that To be sure, it's longer this way.
riage of equations to readable prose. Fisher's rule is for the benefit not of Consistent use of the Good Samaritan
I present a few of them here, empha- the author, but the reader. rule might well increase the length of
sizing that the list makes no claim to For although you, dear author, may your paper by a few percent. But
be complete. We are constantly as- have no need to refer in your text to admit it. Your paper is probably
the equations you therefore left un- already too long by at least 30%
numbered, it is presumptuous to as- because you were in such a rush to get
David Mermin is a professor at Cornell sume the same disposition in your it out that you didn't really take
University and is now happily in his readers. And though you may well enough care putting it all together.
final year as director of the Laboratory have acquired the solipsistic habit of So prune elsewhere, but don't force
of Atomic and Solid-State Physics. He writing under the assumption that your poor readers—you really must
has worked in low-temperature physics, you will have no readers at all, you assume you will have some, or it is
statistical physics, foundations of are wrong. There is always the ref- madness to go on writing—to go
quantum mechanics and eree. The referee may desire to make leafing back when a few words from
quasicrystallography. reference to equations that you did you would save them the trouble.

© 1989 American \aa PHYSICS TODAY OCTOBER 1989 9

 REFERENCE FRAME
Admittedly sometimes an equation should end with a period or, rarely, if above signal the start of a new and
is buried so deep in the guts of an the equation terminates an interroga- better tradition.
argument, so contingent on context, tive sentence, it should end with a We should strive, more generally, to
so ungainly in form that no brief question mark. (Having now succeed- make errant journals mend their
phrase can convey to a reader even a ed in publishing an equation requir- ways. It is easier than you might
glimmer of what it is about, and ing a quotation mark, it remains my think. One of my students and I once
anybody wanting to know why it was dream to publish an article with an did a piece of work that required us to
invoked a dozen pages further on equation that requires a question lead the reader (or at least, we know
cannot do better than to wander back mark; somehow I haven't got around for a fact, the referee) through un-
along the trail and gaze at the equa- to it.) If the equation terminates a avoidably dense thickets of equations.
tion itself, all glowering and menac- clause or is part of an extended list, Unfortunately the otherwise obvious
ing in its lair. Even here, the mere then it should end with a comma or journal for our paper systematically
attempt to apply the Good Samaritan semicolon. Only infrequently is no violated the Math Is Prose rule, so in
rule can have its benefits. If the punctuation required, as, for exam- our letter of submission we empha-
nature of the equation is inherent- ple, in "Only when sized that the punctuation in our
ly uncharacterizable in a compact N equations was essential for the com-
phrase, is the cross-reference really £ f(x,) = 0 (2) prehensibility of our argument. The
necessary? Indeed, is the equation letter of acceptance, however, in-
itself essential? Or is it the kind of is it impermissible to divide by this formed us that the publisher adhered
nasty and fundamentally uninterest- sum." in this and all its other journals, as
ing intermediate step that readers We punctuate equations because well as in its books, to a firm policy of
would either skip over or, if seriously they are a form of prose (they can, never punctuating equations. In that
interested, work out for themselves, after all, be read aloud as a sequence case, we wrote back, just return the
in neither case needing to have it of words) and are therefore subject to manuscript and we'll send it some-
appear in your text? If so, drop it. the same rules as any other prose. To where else. After a long pause we
You will then have to revise the decree that every sentence should end were informed that at a meeting of the
argument that referred back to it, but in a period unless the sentence ter- board of directors of the publishing
the chances are good that the argu- minates in a displayed equation is firm a special dispensation had been
ment will gain in clarity from not grotesque. (If you disagree, try the granted to our paper, and indeed, it
having an uncharacterizable monster rule that every opening quotation appeared with punctuated equations.1
of an equation at its heart. mark must be followed by a closing Fortunately Fisher's rule and the
Rule 3 (Math Is Prose rule). The one unless the quotation terminates Good Samaritan rule don't require
Math Is Prose rule simply says: End in an equation.) But one does not assent from boards of directors, so you
a displayed equation with a punctua- punctuate equations only because it is have nobody to blame but yourself if
tion mark. It is implicit in this ugly not to; more importantly, punc- your papers don't observe them; you
statement that the absence of a punc- tuation makes them easier to read can mend your ways right now. At a
tuation mark is itself a degenerate and often clarifies the discussion in minimum you will make life much
form of punctuation that, like periods, which they occur. Acquiring the hab- easier for an overworked referee, and
commas or semicolons, can be used it of viewing an equation not as a with luck you might even have a few
provided it makes sense. For unlike grammatically irrelevant blob, but as happily undistressed readers.
the figures and tables in your article, a part of the text fully deserving of
unlike droppings on a lawn, the equa- punctuation, can only improve the Reference
tions you display are embedded in fluency and grace of one's expository 1. A. Garg, N. D. Mermin, Found. Phys.
your prose and constitute an insepa- mathematical prose. 14, 1 (1984). •
rable part of it. The detailed theory of Most journals punctu-
how equations are to be viewed as ate their equations, even
prose need not concern us here. if the author of the manu-
Sometimes they function as subordi- script did not, but a sorry
nate clauses, the equals sign being the few don't, removing all
verb; sometimes they appear as sub- vestiges of the punctua-
stantive phrases, like a list of the tion carefully supplied by
contents of a room; sometimes, regret- the author. This unavoid-
tably, they must merely be presented ably weakens the cou-
to the reader as objects like quota- pling between the math
tions (but with the convention that and the prose, and often
quotation marks are not required introduces ambiguity and
[except in the rare case that Math Is confusion. I'm sorry to
Prose requires it, as in Eq. (1) above say that PHYSICS TODAY is
(which I never dreamed I would be guilty of this practice. To
referring back to when I first put it be sure, its use of equa-
into this essay)]). tions is sufficiently light
Regardless, however, of the often that this does not inflict
subtle question of how to parse the substantial hardship on
equation internally, certain things readers, but it greatly un-
are clear to anyone who understands dermines the role PHYSICS
the equation and the prose in which it TODAY so commendably
is embedded. Thus the end of the plays in other respects as
equation may or may not coincide a model of good writing
with the end of the sentence in which about hard science. May
it stands. If it does, then the equation the appearance of Eq. (1) TiWi> tSrtT AMY N ?RiZ£

PHYSICS TODAY OCTOBER 1 9 8 9 11