An introduction to TEX and friends

Gavin Maltby

November 1992
Contents

1 Getting acquainted with TEX 1

1.1 The spirit of TEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 TEX is a typesetter, not a word-processor . . . . . . . . . . . . . . . 1
1.1.2 Typical TEX interfaces . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Getting started with LaTEX 6

2.1 Why start with LaTEX? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 LaTEX formats, and we compose . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Document styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Preparing a non-mathematical document . . . . . . . . . . . . . . . . . . . 9
2.4.1 Sentences and paragraphs . . . . . . . . . . . . . . . . . . . . . . . 9
2.4.2 Punctuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4.3 Ties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4.4 Specially reserved symbols . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.5 So what are control symbols and words? . . . . . . . . . . . . . . . 14
2.4.6 Commands to change appearance . . . . . . . . . . . . . . . . . . . 16
2.4.7 Accents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4.8 Over-ruling some of TEX’s choices . . . . . . . . . . . . . . . . . . . 18
2.4.9 Commenting your document . . . . . . . . . . . . . . . . . . . . . . 20
2.4.10 Footnotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4.11 Topmatter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4.12 Sectioning commands . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.4.13 LaTEX environments . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.4.14 em environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.4.15 quote and quotation environments . . . . . . . . . . . . . . . . . . 23
2.4.16 verse environment . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.4.17 center environment . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.4.18 flushright and flushleft environments . . . . . . . . . . . . . . 25
2.4.19 verbatim environment . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.4.20 itemize, enumerate, description environments . . . . . . . . . . 27
2.4.21 tabbing environment . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.4.22 tabular environment . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.4.23 figure and table environments . . . . . . . . . . . . . . . . . . . . 33

i
 2.4.24 The letter document style . . . . . . . . . . . . . . . . . . . . . . 34
2.4.25 Common pitfalls; Error messages . . . . . . . . . . . . . . . . . . . 34
2.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3 Mathematical typesetting with LaTEX 39

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2 Displaying a formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3 Using mathematical symbols . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3.1 Symbols available from the keyboard . . . . . . . . . . . . . . . . . 44
3.3.2 Greek letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.3.3 Calligraphic uppercase letters . . . . . . . . . . . . . . . . . . . . . 45
3.3.4 Binary operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3.5 Binary relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.6 Miscellaneous symbols . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3.7 Arrow symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3.8 Expression delimiters
Z . . . . . . . . . . . . . . . . . . . . . . . . . . 47
X
3.3.9 Operators like and . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3.10 Accents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.4 Some common mathematical structures . . . . . . . . . . . . . . . . . . . . 49
3.4.1 Subscripts and superscripts . . . . . . . . . . . . . . . . . . . . . . 49
3.4.2 Primes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.4.3 Fractions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.4.4 Roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.4.5 Ellipsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.4.6 Text within an expression . . . . . . . . . . . . . . . . . . . . . . . 52
3.4.7 Log-like functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.4.8 Over- and Underlining and bracing . . . . . . . . . . . . . . . . . . 53
3.4.9 Stacking symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.4.10 Operators; Sums, Integrals, etc. . . . . . . . . . . . . . . . . . . . . 54
3.4.11 Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.4.12 Changes to spacing . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.5 Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.6 Theorems, Propositions, Lemmas, . . . . . . . . . . . . . . . . . . . . . . . . 57
3.7 Where to from here? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.8 AMS-LaTEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

ii
List of Tables

2.1 Commands for selecting type styles . . . . . . . . . . . . . . . . . . . . . . 17

2.2 LaTEX size-changing commands. . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3 Control sequences for accents . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 LaTEX sectioning commands . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3.1 Lowercase Greek letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

3.2 Uppercase Greek letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.3 Binary Operation Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.4 Binary relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.5 Miscellaneous symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.6 Arrow symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.7 Delimiters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.8 Variable-sized symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.9 Math accents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.10 Log-like functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

iii
Chapter 1

Getting acquainted with TEX

TEX is well known to be the typesetting package, and a vast cult of TEX lovers has evolved.
But to the beginning TEX user, or to someone wondering if they should bother changing
to TEX, it is often not clear what all the fuss is about. After all, are not both WordPerfect
and Ventura Publisher capable of high quality output? Newcomers have often already seen
what TEX is capable of (many books, journals, letters are now prepared with TEX) and so
expect to find a tremendously powerful and friendly package. In fact they do, but that fact
is well hidden in one’s initial TEX experiences. In this chapter we describe a little of what
makes TEX great, and why other packages cannot even begin to compete. Be warned that
a little patience is required—TEX’s virtues are rather subtle to begin with. But when the
penny drops, you will wonder how you ever put up with anything different.

1.1 The spirit of TEX

In order to really appreciate TEX one needs to get a feel for what I call the “spirit” of TEX.
When TEX appears to be making me work overtime to achieve something that I think
ought to be perfectly straightforward, consultation with the TEX spirit shows me the error
of my ways.

1.1.1 TEX is a typesetter, not a word-processor

TEX was designed with no limiting application in mind. It was intended to be able to
prepare practically any document—from a single page all-text letter to a full blown book
with huge numbers of formulae, tables, figures etc. The size and the complexity of a
TEXable document is limited only by hardware considerations. Furthermore, TEX seeks
to achieve all this whilst setting typesetting standards of the highest order for itself. The
expertise of generations of professional printers has been captured in TEX, and it has been
taught all the tricks of the trade.
Historically, printers prepared a document by placing metal characters in a large tray
and arranging and binding them to form a page. This was very precisely done, but the

1
ultimate precision was limited because of the mechanical nature of things and by time
considerations. TEX prepares a page in an analogous manner (putting your characters
and formulae into “boxes” which are then “glued” together to form the page), but has
the advantage of enormous precision because placement calculations are performed by
computer. Indeed, TEX’s internal unit (the “scaled point”) is about one-hundredth of the
wavelength of natural light!
“But conventional word processors run on computers , too”, you object. Yes, but
their fundamental limitation is that they try to “keep up” with you and “typeset” your
document as you type. This means that it can only make decisions at a local level (eg, it
decides where to break a line just as you type the end of the line). TEX’s secret is that
it waits until you have typed the whole document before it typesets a single thing! This
means that TEX can make decisions of a global nature in order to optimise the aesthetic
appeal of your document. It has been taught what looks good and what looks bad (having
been given a measure of the “badness” of various possibilities) and makes choices for your
document that are designed to make it “minimally bad”.
But TEX’s virtues run much deeper than that, which is just as well because it is possible
to get satisfactory, though imperfect, results from some word processors. One of TEX’s
strongest points is its ability to typeset complicated formulae with ease. Not only does
TEX make hundreds of special symbols easily accessible, it will lay them out for you in
your formulae. It has been taught all the spacing, size, font, . . . conventions that printers
have decided look best in typeset formulae. Although, of course, it doesn’t understand any
mathematics it knows the grammar of mathematics—it recognises binary relations, binary
operators, unary operators, etc. and has been taught how these parts should be set. It is
consequently rather difficult to get an equation to look bad in TEX.
Another advantage of compiling a document after it is typed is that cross-referencing
can be done. You can label and refer back to chapters, sections, tables etc. by name rather
than absolute number, and TEX will number and cross-reference these for you. Similarly,
it will compile a table of contents, glossary, index and bibliography for you.
Essential to the spirit of TEX is that it formats the document whilst you just take care
of the content, making for increased productivity. The cross-referencing just mentioned
is just part of this. Many more labour-saving mechanisms are provided for through style
files. These are generic descriptions of classes of documents, teaching TEX just how each
class likes to be formatted. This is taught in terms of font preferences, default page sizes,
placement of title, author, date, etc. For instance, a paper style file could teach TEX
that when typesetting a theorem it should embolden the part that states the theorem
number and typeset the text of the theorem statement in slanted Roman typeface (as
in many journals). The typist simply provides and indication that a theorem is being
stated, and then types the text of the theorem without bothering to choose any fonts or
do any formatting—all that is done by the style file. Style files exist for all manner of
document—letters, articles, papers, books, proceedings, review articles, and so on.
In addition to style files, there are macro packages. A macro is just a definition of a
new TEX command in terms of existing ones. Don’t think small when you think of macros!
When typing a document that has a lot of repetition in it, say the same expression is used

2
again and again in different different equations, you can define a macro in your document
to abbreviate that expression. But macros can teach TEX how to typeset all sorts of
complicated structures, not just parts of an equation. Many macro packages (files that
are just collections of definitions) have been written to teach TEX all sorts of applications.
There are specialist maths packages (AMS-TEX, AMS-LaTEX), general purpose packages
(LaTEX), packages for setting tree diagrams, Feynmann diagrams, languages like Chinese,
Arabic and Ancient Greek, orchestral scores, and many, many more. All these are freely
available, a spin-off of the giant TEX cult.
Another facet of the design of TEX allows it to use practically any output device. In
fact, TEX doesn’t talk to any printers, screens, phototypesetters at all! Instead, when a
document is compiled a device independent (.dvi) is produced—TEX does not compile
with any particular output device in mind. Printer drivers are then invoked on this .dvi
file and, in consultation with the font data for that printer, produce output suitable for
the particular device. You can choose an HP Laserjet driver, or an Apple LaserWriter
driver, or a dot matrix driver etc. All use the same .dvi file as input (and remember
the material in there is set to enormous accuracy) and attempt to image that file on the
particular device as faithfully as possible. If you are using a top of the line laser printer or
phototypesetter, then TEX’s massive internal precision will not be wasted. Alternatively,
a dot matrix printer will give a coarse approximation of the ideal image that is suitable
only for proof-reading. In addition to portability, these .dvi files help ensure that there
are very few printing surprises when you move from one device to another: how many
times has your favourite word-processor made you reformat a document when you wish to
change printers?
There are many other motivations one could cite for the superiority of TEX. But it is
time that we started to get our hands dirty. One last comment: TEX was not designed to
supplant secretaries and professional printers—it was designed to aid them in their work
and, in the words of the TEX designer Donald Knuth, allow them to “go forward and
create masterpieces of the publishing art”.It also allows those who generate the material
to be typeset—mathematicians, physicists, computer scientists, etc—to prepare their own
documents in a language that is intimately linked to the language we use for writing such
material.
The novice reader will still have no idea of what a TEX source file looks like. Indeed,
why do we keep referring to it as a source file? The fact of the matter is that TEX is
essentially a programming language. Just as in any compiled language (e.g., Pascal, C)
one prepares a source file and submits it to the compiler which attempts to produce an
object file (.dvi file in the TEX case). To learn TEX is to learn the command syntax of
the commands that can be used in the source file.

1.1.2 Typical TEX interfaces

TEX was designed to run on a multitude of computers. It is therefore the case that the
documentation for TEX and its “friends” LaTEX, AMS-TEX, etc. is not computer specific.
Only command syntax is described—i.e., the content of your source file—but few details

3
of how to get from there to a printout are given. Those details are left to site-specific
documents.
The average user loses little in using TEX on, say, a PC rather than on a bigger machine.
Indeed, compilation times on the new PCs begin to rival those on a Sun Sparc Station 2 (no
slouch). Running on top of DOS can cause memory problems when very large documents
are being prepared. That aside, the quality of the document is not affected because of
the careful design of TEX—whether you work on a a machine with massive floating point
precision or a modest XT the .dvi files produced on compilation will be identical; and
when those files are submitted to printer equivalent printer drivers (say for an HP LaserJet
III attached to a Sun in one case and a PC in the other) the output will be identical
because the font information they draw on is identical.
By the nature of TEX most time is spent editing the source document (before submitting
it for compilation). No special interface is necessary here, you just use your favourite text
editor (perhaps customising it to enhance TEXnical typing. Thus TEX user interfaces are
usually small and simple, often even missing. One frequently uses TEX at command line
level, just running the editor, compiler etc. as you need them. Sometimes a TEXshell
program is present, which runs these for you when you choose various menu options.
Whatever the interface, there are just a few basic steps to preparing a document:

1. Choose a document style to base your document on (e.g., letter, article).

2. Glance through the material you have to type, and decide what definitions might be
made to save you a lot of time. Also, decide on the overall structure of the prospective
document (e.g., will the largest sectional unit be a chapter or a part?). If you are
going to compose as you type, then pause a moment to think ahead and plan the
structure of your document. The importance of this step cannot be overstressed, for
it makes clear in your mind what you want from TEX.

3. Prepare your input file, specifying only the content and the logical structure (parts,
sections, theorems,...) thereof and forgetting about formatting details.

4. Submit your input, or source, file to the TEX compiler for compilation of a .dvi file.

5. If the compiler finds anything in your source file strongly objectionable, say incorrect
command syntax, then return to editing.

6. Run a previewer to preview your compiled document on the screen. Resolution is

only limited by your screen, and can be very good indeed on some modern monitors.

7. Go back to editing your document until glaring errors have been taken care of.

8. Make a printout of your compiled document, and check for those errors that you
failed to notice on the screen.

4
Performing these steps may be effected through typing at the system prompt (barebones
technique) or through choosing menu options in a TEXshell program. The latter will
probably provide some conveniences to make your life easier.
If you think this sounds like a lot of work, it is time that you consult with the TEX
spirit! Sure your first couple of tries may be hesitant, but before long you’ll find that you
can take less time to prepare a document on TEX than on any other package.

5
Chapter 2

Getting started with LaTEX

2.1 Why start with LaTEX?

To answer this question we must say a little more about some of the macro packages we
mentioned earlier.
The TEX typesetting system was designed by the eminent Stanford computer scientist
Donald Knuth, on commission from the American Mathematical Society. It was designed
with enormous care, to be ultimately powerful and maximally flexible. The enormous
success of Knuth’s design is apparent from the vast number of diverse applications TEX
has found. In reading the following you must keep one thing clearly in mind: there is
only TEX language, and all the other packages whose names end in the suffix -TEX simply
harness it’s power via a whole lot of complicated macro definitions.
TEX proper is a collection of around 300 so called primitive typesetting commands.
These work at the very lowest level, affording enormous power. But to make this raw power
manageable, some macros must be defined to tame raw TEX somewhat. The standard set
of macros is called Plain TEX, and consists of about 600 macro definitions. It is clear that
these definitions must be made in terms of TEX primitives, or in terms of previously made
definitions. Plain TEX, however, is still no place for the timid. A strong working knowledge
of TEX is still required to understand the ins and outs of Plain TEX.
In the few years after the initial TEX release (1982), the macro packages AMS-TEX
and LaTEX were born. AMS-TEX was written by Michael Spivak, also on commission from
the AMS. This package was designed to facilitate the preparation of the numerous books,
journals, and review indices that fall under the auspices of the AMS and its affiliates. Mar-
ried to the macro package was a style file—the AMS preprint style. This was distributed
along with the macro package, so that authors submitting to journals could use it in the
preparation of their articles. The given style was based on the style used by the Journal of
the American Mathematical Society, i.e., it conformed to their page sizes and typographical
conventions. This meant that people around the world produced papers that were all based
on the same style. The clever part is this: when a source file is submitted to a journal other
than the Journal of the AMS, the journal staff simply substitute their style file for the AMS

6
preprint style and the paper will appear completely different with no other changes to the
source code! To create their style file, a journal just needed to tweak the standard AMS
prepint style: for instance, the original preprint style places author addresses at the very
end of a paper; If a journal wishes this to appear on the first page then they just modify
their in-house version of the style file, and the change will be effected without having to
change the file submitted by the author.
LaTEX was written for more general usage. It lacks some of the mathematical finesse
inherited by AMS-TEX from the vast experience of the AMS technical staff, but more than
makes up for this in its ability to enhance the typesetting of letters, books, poetry, etc.
LaTEX also scores high points for its enhanced command syntax.
With AMS-TEX and LaTEX being released at around the same time (1984–1985), there
were born many AMS-TEX literate but LaTEX illiterate users, and conversely. LaTEX was
easier to learn because of its more friendly syntax, and also provided powerful cross-
referencing commands that AMS-TEX did not. So the AMS commissioned another project
to furnish LaTEX users with the additional power of AMS-TEX while not compromising the
LaTEX command syntax or cross-referencing commands. This resulted in the AMS-LaTEX
macro package and associated style file for submission to journals.
That is why we will kick off our TEX careers with LaTEX! It is easier to learn and provides
many conveniences, and the user who requires additional mathematical typesetting prowess
can easily move on to AMS-LaTEX. Much of what we say will be true for TEX itself, but we
shall regard LaTEX as the lowest common-denominator. By far the majority of LaTEX and
AMS-LaTEX users will never have to learn “raw” TEX, for they will be shielded from direct
exposure by the numerous powerful macro packages. In the rare case that something way
out of the ordinary is required, the local TEX guru can be consulted.

2.2 LaTEX formats, and we compose

The free form nature of the input file is essential to the spirit of TEX. As we type, we do
not concern ourselves with linebreaks and pagebreaks so much as the content of what we
are typing. In fact, we’ll see that TEX will choose nice line breaks even for bizarre looking
input. This is just part of the concept of only having to describe the logical structure of
the document to LaTEX, and not worry about nuisance-value formatting details. We inform
LaTEX of the logical structure of our document by telling it when to begin a new paragraph,
subsection, section, chapter, theorem, definition, remark, poem, list etc. When typing a
particular element of the logical structure, we need pay little attention to how we lay our
source file out.
A consequence of this is that we have to go to a bit of effort to mess things up. Starting
a new line, for instance, entails more than just pressing Return because LaTEX will just
regard the next word you type as exactly that—the next word in the paragraph. You have
to specifically ask for a line to be terminated. Things like this may seem to be a bit of a
nuisance, but it is a small price to pay for the automatic formatting that necessitated it.
Further, such small inconveniences have been localised to rare events. I have, for instance,

7
 not once forced a new line up until this point in the present document.

2.3 Document styles

We have explained the concept of a document style during our discussion of the virtues of
TEX and the discussion of AMS-TEX. It remains to name a few, and indicate where they
would be used. One always has to choose a document style when preparing a document
with LaTEX.
The basic document styles in LaTEX are letter, article, report, and book. Many
more are available, but these few cover the majority of straightforward applications. This
is because styles are not rigid—you can impose your own parameter choices if you want. So
one chooses the style that most closely approximates the document you have in mind, and
performs some minor tweaks here and there. The article style is used for documents that
are to have the appearance of a journal or magazine article. The report style is usually
used for larger documents than the article style. These styles really only differ in their
choice of default page size, font, placement of title and author, sectional units, etc. and on
how they format certain LaTEX constructs. You use the same LaTEX commands in each.
Since the examples here will be small, we will choose to use the article document style.
There are a number of possible options with each document style. The syntax for
choosing a document style follows. Don’t worry if this leaves you with no idea of how
to choose a document style, for we will soon be seeing some examples. Also, remember
that an argument in square brackets is optional, and can omitted altogether (including the
brackets). \documentstyle [options]{style} where style is the main document
style (eg report) and the optional argument options is a list of document style options
chosen ¿from the following list:

11pt chooses 11-point as the default font size for the document, instead of the default
10-point.

12pt chooses 12-point as the default font size.

twoside formats output as left and right pages, as in a book.

twocolumn produces two-column magazine like output.

titlepage applies to the article style only, causing the title and abstract to appear on a page
each.

In fact there are many, many more document style options but we won’t mention any
more here.

8
2.4 Preparing a non-mathematical document
We assume that you have read the local guides to TEX at your site and have decided which
system environment you want to work in. There you have been shown how to perform the
steps required to produce a printed document from a LaTEX source file.

2.4.1 Sentences and paragraphs

Let’s create our very first LaTEX document, which will consist of just a few paragraphs.
As mentioned above, paragraph input is free-form. You type the words and separate
them by spaces so that LaTEX can distinguish between words. For these purposes, pressing
Return is equivalent to inserting a space—it does not indicate the end of a line, but the
end of a word. You tell LaTEX that a sentence has ended by typing a period followed by
a space. LaTEX ignores extra spaces; typing three or three thousand will get you no more
space between the words that these spaces separate than typing just one space. Finally,
you tell LaTEX that a paragraph has ended by leaving one or more blank lines. In summary:
LaTEX concerns itself only with the logical concepts end-of-word, end-of-sentence, and end-
of-paragraph. Sounds complicated? An example should clear things up:
\documentstyle{article}
\begin{document}
Words within a sentence are ended by spaces. One space
between words is equivalent to any number. We are only
interested in separating one word from the
next, not in formatting the space between them.
For these purposes, pressing Return
at the end of a line
and starting a new word on the next line
just serves to separate
words, not to cut a line short.
The end of a sentence is indicated by a period
followed by one or more spaces.

The end of a paragraph is indicated by leaving a blank line.

All this
means that we can type without too much regard for layout, and
the typesetter will sort things out for us.
\end{document}

produces the result

9
 Words within a sentence are ended by spaces. One space between words is equiv-
alent to any number. We are only interested in separating one word from the next,
not in formatting the space between them. For these purposes, pressing Return at
the end of a line and starting a new word on the next line just serves to separate
words, not to cut a line short. The end of a sentence is indicated by a period followed
by one or more spaces.
The end of a paragraph is indicated by leaving a blank line. All this means that
we can type without too much regard for layout, and the typesetter will sort things
out for us.
Perhaps you would like to try running LaTEX on the above input. Consult your local
guide for details.
Note that we have learned more than just how LaTEX recognises words, sentences and
paragraphs. We’ve also seen how to specify our choice of document style and how to tell
LaTEX where our document begins and ends. Any material that is to be printed must lie
somewhere between the declaration of \begin{document} and that of \end{document}.
Definitions that are to apply to the entire document can be made before the declaration
of the document beginning. The specification of document style must precede all other
material.
In future examples we won’t explicitly display the commands that select document style
and delimit the start and end of the document. But if you wish to try any of the examples,
don’t forget to include those commands. The article document style will do for most of
our examples. Of course, the preceding example looks not at all like an article because it
is so short and because we specified no title or author information.
Most of what you need to know to type regular text is contained in the example above.
When you consider that by far the majority of any document consists of straight text, it
is obvious that LaTEX makes this fabulously straightforward. LaTEX will do all the routine
work of formatting, and we simply get on with the business of composing.
LaTEX does more than simply choose pleasing line breaks and provide natural spacing
when setting a paragraph. Remember we said that TEX has inherited the knowledge of
generations of professional printers—well part of that knowledge includes being on the
look-out for ligatures. These are combinations of letters within words which should be
typeset as a single special symbol because they will “clash” with each if this is not done.
Have a look at these words
flight, flagstaff, chaff, fixation
and compare them with these
flight, flagstaff, chaff, fixation
See the difference? In the first set I let LaTEX run as it usually does. In the second I
overruled it somewhat, and stopped it from creating ligatures. Notice how the ‘fl’, ‘ff’, and
‘fi’ combinations are different in the two sets—in the former they form a single symbol (a
ligature) and in the latter they are comprised of two disjoint symbols. There are other
combinations that yields ligatures, but we don’t have to bother remembering any of them
because LaTEX will take care of these, too.

10
 Notice, too, that LaTEX has been taught how to hyphenate the majority of words. It
will hyphenate a word if it feels that the overall quality of the paragraph will be improved.
For long words it has been taught several potential hyphenation positions.
LaTEX also goes to a lot of trouble to try to choose pleasing page breaks. It avoids
“widows”, which are single lines of a paragraph occurring by themselves at either the
bottom of a page (where it would have to be the first line of a paragraph) or at the top of
a page (where it would have to be the last). It also “vertically justifies” your page so that
all pages have exactly the same height, no matter what appears on them. As testimony to
the success of the pagebreaking algorithm, I have (to this point) not once chosen a page
break in this document.

2.4.2 Punctuation
Typists have a convention whereby a single space is left after a mid-sentence comma, and
two spaces are left after a sentence-ending period. How do we enforce this if LaTEX treats
a string of spaces just like a single one? The answer, unsurprisingly, is that we don’t.
To have a comma followed by the appropriate space, we simply
type a comma follows by at least one space. To end a sentence
we type a period with at least one following space. No space will
be inserted if we type a comma or period followed straight away
by something other than a space, because there are times when
we won’t require any space, i.e., we do what comes narurally.

will produce
To have a comma followed by the appropriate space, we simply type a comma
follows by at least one space. To end a sentence we type a period with at least one
following space. No space will be inserted if we type a comma or period followed
straight away by something other than a space, because there are times when we
won’t require any space, i.e., we do what comes naturally.
LaTEX will produce suitable space after commas, periods, semi-colons and colons, excla-
mation marks, question marks etc. if they are followed by a space. In stretching a line to
justify to the right margin, it also knows that space after a punctuation character should
be more “stretchable” than normal inter-word space and that space after a sentence-ending
period should be stretched more than space after a mid-sentence comma. TEX knows the
nature of punctuation if you stick to the simple rules outlined here. As we’ve already said,
those rules tell LaTEX how to distinguish consecutive words, sentences, phrases, etc.
Actually, there is more to ending sentences than mentioned above. Since LaTEX cannot
speak English, it works on the assumption that a period followed by a space ends a sentence
unless the period follows a capital letter. This works most of the time, but can fail. To get
a normal inter-word space after a period that doesn’t end a sentence, follow the period by
a control space—a \ (a \ character followed by a space or return). Very rarely, you will
have to force a sentence to end after a period that follows a capital letter (remember that
LaTEX assumes this doesn’t end a sentence). This is done by preceding the period with a
\@ command (you can guess from the odd syntax that this is rarely needed).

11
 It’s time we saw some examples of this. After all, this is our first experience of control
symbols (don’t worry, there are many more to come).
We must be careful not to confuse intra-sentence periods
with periods that end a sentence, i.e.\ we must remember
that our task is to describe the sentence structure. Periods
that the typesetter requires a little help with typically result
>from abbreviations, as in etc.\ and others. We have to work
somewhat harder to break a sentence after a capital letter,
but that shouldn’t bother us to much if we keep up our intake
of vitamin E\@. All this goes for other sentence-ending
punctuation characters, so I could have said vitamin E\@!
Fortunately, these are rare occurrences.

results in
We must be careful not to confuse intra-sentence periods with periods that end a
sentence, i.e. we must remember that our task is to describe the sentence structure.
Periods that the typesetter requires a little help with typically result ¿from abbrevi-
ations, as in etc. and others. We have to work somewhat harder to break a sentence
after a capital letter, but that shouldn’t bother us to much if we keep up our intake
of vitamin E. All this goes for other sentence-ending punctuation characters, so I
could have said vitamin E! Fortunately, these are rare occurrences.
Quotation marks is another area where LaTEX will do some work for us. Keyboards
have the characters ‘, ’, and " but we want to to have access to each of ‘, ’, “, and ”. So
we proceed like this:
‘\LaTeX’ is no conventional word-processor, and
to to get quotes, like ‘‘this’’, we type repeated
‘ and ’ characters. Note that modern
convention is that ‘‘punctuation comes after
the closing quote character’’.

which gives just what we want

‘LaTEX’ is no conventional word-processor, and to to get quotes, like “this”, we
type repeated ‘ and ’ characters. Note that modern convention is that “punctuation
comes after the closing quote character”.
Very rarely, you have three quote characters together. Merely typing those three quote
characters one-after-the-other is ambiguous—how should they be grouped? You tell LaTEX
how you want them grouped by inserting a very small space called \,.
‘‘\,‘Green ham’ or ‘Eggs?’\,’’ is the question.

gives the desired result

“ ‘Green ham’ or ‘Eggs?’ ” is the question.
Since we have a typesetter at our disposal, we might as well use the correct dashes
where we need them. There are three types of dash: the hyphen, the endash, and the
emdash. A minus sign is not a dash.

12
 Hyphens are typed as you’d hope, just by typing a - at the point in the word that
you want a hyphen. Don’t forget that LaTEX takes care of hyphenation that is required
to produce pretty linebreaks. You only type a hyphen when you explicitly want one to
appear, as in a combination like “inter-college”.
An endash is the correct dash to use in indicating number ranges, as in “questions 1–3”.
To specify an endash you type two consecutive dashes on the keyboard, as in 1--3.
An emdash is a punctuation dash, used at the end of a sentence—I tend to use them
too much. To specify an emdash you type three consecutive dashes on the keyboard, as in
“. . . a sentence---I tend to. . . ”.
Theorems 1--3 concern the semi-completeness
of our new construct---in the case that it
satisfies the first axiom.

yields
Theorems 1–3 concern the semi-completeness of our new construct—in the case
that it satisfies the first axiom.

2.4.3 Ties
When you always remember to use ties, you know that you are becoming TEXnically
inclined. Ties are used to prevent LaTEX from breaking lines at certain places. LaTEX will
always choose line breaks that result in the most aesthetically pleasing paragraph as judged
by its stringent rules. But because LaTEX does not actually understand the material it is
setting so beautifully, it can make some poor choices.
A tie is the character ~. It behaves as a normal interword space in all respects except
that the line-breaking algorithm will never break a line at that point. Thus
Dr. Seuss should be typed as Dr.~Seuss

for this makes sure that LaTEX will never leave the ‘Dr’ at the end of one line and put the
‘Seuss’ at the beginning of the next.
One should try to get in to the habit of typing ties first-time, not after waiting to see
if LaTEX will make a poor choice. This will allow you to make all sorts of changes to your
text without ever having to go back and insert a tie at a point that has migrated to the
end of a line from the middle of a line as a result of those changes. Remember, of course,
that the line-breaks
Here are some more examples of places where you should remember to place ties.

13
 Chapter~10 Donald~E. Knuth
Appendix~C width~2
Figure~1 function~f
Theorem~2 1,~2, or~3
Lemmas 3 and~4 equals~5

2.4.4 Specially reserved symbols

In the sequel we will see that the the ten characters
# $ % & ~ _ ^ \ { }
are reserved for special use. Indeed, we have seen already that \ and ~ are non-printing
characters that perform special services (and we’ll have a lot more to say about the use of
\).
But what if we need one of these special symbols to appear in our document? The
answer for seven of the symbols is to precede them by a \ character, so forming another
control symbol (remember that \ followed by a space was also a control symbol).
It is not 100\% straightforward to typeset the
characters \$ \& \% \_ \{ \}, but given the
enormous convenience of the use they are normally
reserved for this is a small price to pay.

produces
It is not 100% straightforward to typeset the characters $ & % { }, but given
the enormous convenience of the use they are normally reserved for this is a small
price to pay.

2.4.5 So what are control symbols and words?

In typing a document, we can think of ourselves as being in one of two distinct modes. We
are either typing literal text (which will just be set into neat paragraphs for us) or we are
typing text that will be interpreted by LaTEX as an instruction to insert a special symbol
or to perform some action. Thus we are either typing material that will go straight into
the document (with some beautification), or we are giving commands to LaTEX.
Some commands are implicit, in that we don’t have to do anything much extra. For
instance, we command LaTEX to end the present sentence by typing a period (that doesn’t
follow a capital letter). These are not so much commands as part of having to describe
the logical structure of a document.
A control word is something of the form \commandname, where the command name is
a word made up only of the letters a to z and A to Z. A control symbol consists of a \
followed by single symbol that is not a letter.
Here are some examples:

14
 • we have met the control space symbol \ before,

• the commands to set symbols like % and $ are control symbols

• \@ was a control symbol that told told LaTEX that the very next period did really
end the sentence,

• \LaTeX is a control word that tell LaTEX to insert its own name at the current point,

• \clubsuit instructs that a ♣ be inserted,

• \pounds inserts a £ symbol,

• \S inserts a ¶ symbol,

• \em makes the ensuing text be emphasised,

These examples show that control sequences can be used to access symbols not available
from the keyboard, do some typesetting tricks like setting the word LaTEX the way it does,
and change the appearance of whole chunks of text as with \em. We’ll be meeting many
more of these type of control sequences.
Another enormously powerful class of control sequences is those that accept arguments.
They tell LaTEX to take the parts of text you supply and do something with them—like
make a fraction by setting the first argument over the second and drawing a line of the
appropriate length between them. These are part of what makes LaTEX so powerful, and
here are some examples.

• \chapter{The beginning} causes LaTEX to start a new chapter with name “The
Beginning”, number it in sequence, typeset the chapter heading in a suitable font,
and make an entry in the table of contents,

• \overline{words} causes words to be overlined,

• \frac{a+b}{c+d} sets the given two argument as a fraction, doing most of the dirty
work for us: a+b
c+d
,
√
• \sqrt[5]{a+b} typesets the fifth-root of a + b, like this: 5 a + b. The 5 is in square
brackets instead of braces because it is an optional argument and could be ommited
all together (giving the default of square root),

Mandatory arguments are given enclosed by braces, and optional arguments enclosed
by square brackets. Each command knows how many arguments to expect, so you don’t
have to provide any indication of that.
We have actually jumped the gun a little. The above examples include examples of
mathematical typesetting, and we haven’t yet seen how to tell LaTEX that it is typesetting
maths as opposed to some other random string of symbols that it doesn’t understand
either. We’ll come to mathematical typesetting in good time.

15
 We need to dwell on a TEXnicality for a moment. How does LaTEX know where the
name of a control sequence ends? Will it accept both \pounds3 and \pounds 3 in order to
set £3, and will \emWalrus and \em Walrus both be acceptable in order to get Walrus?
The answer is easy when you remember that a control word consists only of alphabetic
characters, and a control symbol consists of exactly one nonalphabetic character.
So to determine which control sequence you typed, LaTEX does the following:

1. when a \ character is encountered, LaTEX knows that either a control symbol or a

control word will follow.

2. If the \ is followed by a nonalphabetic character, then LaTEX knows that it is a control

symbol that you have typed. It then recognises which one it was, typesets it, and
goes on to read the character which follows the symbol you typed.

3. If the \ is followed by an alphabetic character, then LaTEX knows that it is a control

word that you have typed. But it has to work out where the name of the control word
ends and where the ensuing text takes over again. Since only alphabetic characters are
allowed, LaTEX reads everything up to just before that first nonalphabetic character
as the control sequence name. Since it is common to delimit the end of a control
word by a space, LaTEX will ignore any space that follows a control word, since you
want that space treated as end-of-control-word space rather than interword space.

This has one important consequence: The character in the input file immediately after
a control symbol will be “seen” by LaTEX, but any space following a control word will be
discarded and never processed. This does not affect one much if you adopt the convention
of always typing a space after a control sequence name.
There is a rare circumstance where this necessitates a little extra work and thought,
which we illustrate by example:
If we type a control word like \LaTeX in the running text
then we must be cautious, because the string of spaces that
come after it will be discarded by the \LaTeX\ system.

which produces the output

If we type a control word like LaTEXin the running text then we must be cautious,
because the string of spaces that come after it will be discarded by the LaTEX system.

2.4.6 Commands to change appearance

We’ve seen a little of how to access various symbols using control sequences and we men-
tioned the \em command to emphasise text, but we didn’t see how to use them. We look
here at commands that change the appearance of the text.
Each of the control words here is a directive rather than a control sequence that accepts
an argument. This is because potential arguments consisting of text that wants to be

16
emboldened or emphasised are very large, and it would be a nuisance to have to enclose
such an argument in argument-enclosing braces.
To delimit the area of text over which one of these commands has effect (its scope) we
make that text into what is called a group. Groups are used extensively in LaTEX to keep
effects local to an area, rather than affecting the whole document. Apart from enhancing
usability, this also in a sense protects distinct parts of a document from each other.
The LaTEX commands for changing type style are given in table 2.1, and those for
changing type size are given in table 2.2. Commands for selecting fonts other than these
are not discussed here.
\rm Roman \it italic \sc Capitals
\em Emphasised \sl slanted \tt typewriter
\bf boldface \sf sans serif
Table 2.1: Commands for selecting type styles

Each of the type style selection commands selects the specified style but does not change
the size of font being used. The default type style is roman (you are reading a roman style
font now). To change type size you issue one of the type size changing commands in
table 2.2, which will select the indicated size in the currently active style. The release of
LaTEX 3.0 (the present version is 2.09) will see the New Font Selection Scheme in place as
a standard feature. This makes font matters much easier to deal with.
size default (10pt) 11pt option 12pt option
\tiny 5pt 6pt 6pt
\scriptsize 7pt 8pt 8pt
\footnotesize 8pt 9pt 10pt
\small 9pt 10pt 11pt
\normalsize 10pt 11pt 12pt
\large 12pt 12pt 14pt
\Large 14pt 14pt 17pt
\LARGE 17pt 17pt 20pt
\huge 20pt 20pt 25pt
\Huge 25pt 25pt 25pt
Table 2.2: LaTEX size-changing commands.

The point-size option referred to in table 2.2 is that specified in the \documentstyle
command issued at the beginning of the input file. Through it you select that base (or
default) font for your document to be 10, 11, or 12 point Roman. If no options are specified,
the default is 10-point Roman. The table shows, for instance, that if I issue a \large in this
document for which I chose the 12pt document style option the result will be a 14-point
Roman typeface.
We mentioned that to restrict the scope of a type-changing command we will set the
text to be affected off in a group. Let’s look at an example of this.

17
 When we want to {\em emphasise\/} some text we
use the {\tt em} command, and use grouping to
restrict the scope. We can change font {\large sizes}
in much the same way. We can also obtain {\it italicised},
{\bf emboldened}, {\sc Capitals} and {\sf sans serif} styles.

When we want to emphasise some text we use the em command, and use grouping
to restrict the scope. We can change font sizes in much the same way. We can also
obtain italicised, emboldened, Capitals and sans serif styles.
Notice how clever grouping allows us to do all that without once having to use \rm or
\normalsize.
One more thing slipped into that example—an italic correction \/. This is a very small
amount of additional space that we asked to be inserted to allow for the change from
sloping emphasised text to upright text, because the interword space has been made to
look less substantial from the terminal sloping character. One has to keep an eye open for
circumstances where this is necessary. See the effect of omitting an italic correction after
the emphasised text earlier in this paragraph.
One might expect, by now, that LaTEX would insert an italic correction for us. But
there are enough occasions when it is not wanted, and there is no good rule for LaTEX to
use to decide just when to do it for us. So the italic correction is always left up to the
typist.

2.4.7 Accents
LaTEX provides accents for just about all occasions. They are accessed through a variety
of control symbols and single-letter control worlds which accept a single argument—the
letter to be accented. These control sequences are detailed in table 2.3.
Thus we can produce ó by typing \’{o}, ǎ by typing \v{a}, and Pál Erdös by typing
P\’{a}l Erd\"{o}s. Take special care when accenting an i or a j, for they should lose
their dots when accented. Use the control words \i and \j to produce dotless versions of
these letters. Thus the best way to type to type ĕxı̆gent is \u{e}x\u{\i}gent.

2.4.8 Over-ruling some of TEX’s choices

We’ve seen that ties can be used to stop linebreaks occurring between words. But how can
we stop LaTEX from hyphenating a particular word? More generally, how can we stop it
from splitting any given group of characters across two lines. The answer is to make that
group of characters appears as one solid box , through use of an \mbox command.
For instance, if we wanted to be sure that the word
{\em currentitem\/} is not split across lines
then we should type it as \mbox{\em currentitem}.

18
 \‘{o} ò (grave accent)
\’{o} ó (acute accent)
\^{o} ô (circumflex or “hat”)
\"{o} ö (umlaut or dieresis)
\~{o} õ (tilde or “squiggle”)
\={o} ō (macron or “bar”)
\.{o} ȯ (dot accent)
\u{o} ŏ (breve accent)
\v{o} ǒ (háček or “check”)
\H{o} ő (long Hungarian umlaut)
\t{oo} oo (tie-after accent)
\c{o} o̧ (cedilla accent)
\d{o} o. (dot-under accent)
\b{o} o (bar-under accent)
¯
Table 2.3: Control sequences for accents

If for some reason we wish to break a line

in the middle of nowhere, preventing LaTEX ¿from justifying that line to the prevailing
right margin, then we use the \newline command. One can also use the abbreviated form
\\.
We start with a short line.\newline
And now we continue with the normal
text, remembering that where we press
Return in the input file probably won’t
correspond to a line break in the final
document. More short lines\\
are easy, too.

will produce the line breaks we want

We start with a short line.
And now we continue with the normal text, remembering that where we press Return
in the input file probably won’t correspond to a line break in the final document. More
short lines
are easy, too.
A warning is in order: \newline must only end part of a line that is “already set”.
It cannot be used to add additional space between paragraphs, nor to leave space for a
picture you want to paste in. This is not to be awkward, but is just part of LaTEX holding
up its end of the deal by making you have to specially request additional vertical space.
This prevents unwanted extra space from entering your document.
Later we shall see how to impose our own choice of page size, paragraph indentation,
etc. For now we will continue to accept those declared for us in the document style.

19
2.4.9 Commenting your document
It is handy to be able to make comments to yourself in the source file for a document.
Things like “I must rewrite this section” and “This is version 3 of the document” are
likely. It would also be useful to be able to make the compiler ignore certain parts of the
document at times. For this purpose we can use the % character, for all text on an input
line that is after a % which is not part of an occurrence of the control symbol \% is discarded
by the compiler. Here is an example:
There was a 100\% turnout today,
an all-time record. %perhaps I should check this claim!
%Indeed, there are lots of unsubstantiated claims here!
This made for an extremely productive session.

will yields
There was a 100% turnout today, an all-time record. This made for an extremely
productive session.

2.4.10 Footnotes
Inserting footnotes is easy—LaTEX will position and number them for you. You just indicate
exactly where the footnote marker should go, and provide the text of the footnote. The
footnote text will be placed at the bottom of the present page in a somewhat smaller
font, and be separated from the main text by a short horizontal rule1 to conform with
convention. The footnote in the last line was typed like this:
...rule\footnote{See for yourself! It’s easy ...work.} to conform

No space was typed between the rule and the \footnote, because we want the footnote
marker to appear right next to the last letter of the word.
Multiple footnotes2 are obtained just by using the \footnote command again and
again.

2.4.11 Topmatter
We declare the title and author information using the \title and \author commands,
each of which accept a single argument. Multiple authors are all listed in the argument of
\title, separated by \and’s. The \date command can be used to date a document. After
we have declared each of these, we issue a \maketitle command to have them typeset for
us. In the book and article document styles this will result in a separate page; in the
article style the “top matter” will be placed at the top of the first page. The style files
determine the placement and the choice of font.
1
See for yourself! It’s easy when you don’t have to do any work.
2
Here is another footnote

20
 \title{A Thought for the Day}
\author{Fred Basset \and Horace Hosepipe}
\date{November 1992}
\maketitle

will produce something along the lines of

A Thought for the Day

Fred Basset Horace Hosepipe
November 1992

This topmatter must appear after the \begin{document} and before any other printing
material.

2.4.12 Sectioning commands

As part of our task of describing the logical structure of the document, we must indicate to
LaTEX where to start sectional units. To do this we make use of the sectioning commands
shown in table 2.4.

\part \subsection \paragraph

\chapter \subsubsection \subparagraph
\section
Table 2.4: LaTEX sectioning commands

Each sectioning command accepts a single argument—the section heading that is to

be used. LaTEX will provide the section numbering (and numbering of subsections within
sections, etc.) so there is no need to include any number in the argument. LaTEX will also
take care of whatever spacing is required to set the new logical unit off from the others,
perhaps through a little extra space and using a larger font. It will also start a new page
in the case that a new chapter is begun.
The \part command is used for major subdivisions of substantial documents. The
\paragraph and \subparagraph commands are, unfortunately confusing. They are used
to section off a modest number of paragraphs of text—they don’t start new paragraphs
(remember that that was done by leaving a blank line in our input file). The names were
retained for historical reasons.
It is always a good idea to plan the overall sectional structure of a document in advance,
or at least give it a little thought. Not that it would be difficult to change your mind later
(you could use the global replace feature of an editor, for instance), but so that you have
a good idea of the structure that you have to describe to LaTEX.
The sectioning command that began the present sectional unit of this document was

21
 \subsection{Sectioning commands}

and that was all that was required to get the numbered section name and the table of
contents entry.
There are occasions when you want a heading to have all the appearance of a particular
sectioning command, but shouldn’t be numbered as a section in its own right or produce
a table of contents entry. This can be achieved through using the *-form of the command,
as in \section*{...}. We’ll see that many LaTEX commands have such a *-form which
modify their behaviour slightly.
Not only will LaTEX number your sectional units for you, it will compile a table of
contents too. Just include the command \tableofcontents after the \begin{document}
command and after the topmatter that should precede it.

2.4.13 LaTEX environments

Perhaps the most powerful and convenient concept in the LaTEX syntax is that of an
environment. We will see most of the “heavy” typesetting problems we will encounter can
be best tackled by one or other of the LaTEX environments.
Some environments are used to display a portion of text, i.e. to set it off from the
surrounding text by indenting it. The quote and verse environments are examples of
these. The center environment allows us to centre portions of text, while the flushright
environment sets small portions of text flush against the right margin.
But the true power of LaTEX begins to show itself when we look at environments such as
those that provide facilities for itemised or enumerated lists, complex tabular arrangements,
and for taking care of figure and table positioning and captioning. What we learn here
will also be applicable in typesetting some complicated mathematical arrangements in the
next chapter.
All the environments are begun by a \begin{name} command and ended by an \end{name},
where name is the environment name. These commands also serve as begin-group and end-
group3 markers, so that all commands are local to the present environment—they cannot
affect text outside the environment.
We can also have environment embedded within environment within environment and
so on, limited only by memory available on the computer. We must, however, be careful
to check that each of these nested environments is indeed contained within the one just
outside of it.

2.4.14 em environment
We start with a very simple environment, one which provides an alternative to the \em
command. Remember that \em does not accept an argument; it applies to everything
3
See section 2.4.6

22
within its scope as dictated by the group within which it is used. This can be tricky if we
wish to emphasise a large amount of text, for we may forget a group-delimiting brace and
so upset the entire grouping structure of our document. In cases where we fear this might
happen, we can proceed as follows.
\begin{em}
We must always be careful to match our group-delimiting
braces correctly. If the braces in a document are unevenly
matched then \LaTeX\ will become confused because we will
have, in effect, indicated different scopes than we
intended for commands.
\end{em}

which will give

We must always be careful to match our group-delimiting braces correctly. If the
braces in a document are unevenly matched then LaTEX will become confused because
we will have, in effect, indicated different scopes than we intended for commands.
Although LaTEX doesn’t care too much for how we format our source file, it is obviously
a good idea to lay it out logically and readably anyway. This helps minimise errors as well
as aids in finding them. For this reason I have adopted the convention of always placing
the environment \begin and \end commands on lines by themselves.

2.4.15 quote and quotation environments

This environment can be used to display a part of a sentence or paragraph, or even several
paragraphs, in such a manner that the material stands out from the rest of the text. This
can be used to enhance readability, or to simply emphasise something. Its syntax is simple:
Horace smiled and retaliated:
\begin{quote}
\em You can mock the non-WYSIWYG nature of \TeX\
all you like. You don’t understand that that is
precisely what makes \TeX\ enormously more powerful
than that lame excuse for a typesetter you use.
And I’ll bet that from start to finish of preparing
a document I’m quicker than you are, even if you
do type at twice the speed and have the so-called
advantage of WYSIWYG. In your case, what you see
is {\em all\/} you get!
\end{quote}
and then continued with composing his masterpiece of the
typesetting art.

produces the following typeset material:

23
 Horace smiled and retaliated:
You can mock the non-WYSIWYG nature of TEX all you like. You don’t understand that that
is precisely what makes TEX enormously more powerful than that lame excuse for a typesetter
you use. And I’ll bet that from start to finish of preparing a document I’m quicker than you
are, even if you do type at twice the speed and have the so-called advantage of WYSIWYG.
In your case, what you see is all you get!
and then continued with composing his masterpiece of the typesetting art.
That is a much more readable manner of presenting Horace’s piece of mind than em-
bedding it within a regular paragraph. The quote environment was responsible for the
margins being indented on both sides during the quote. This example has also been used
to show how the commands that begin and end an environment restrict the scope of com-
mands issued within that environment: The \em at the beginning of the quote did not
affect the text following the quote. We have also learned here that if we use \em within al-
ready emphasised text, the result is roman type—and we don’t require an italic correction
here because the final letter of ‘all’ was not sloping to the right.
The quotation environment is used in just the same way as the quote environment
above, but it is intended for setting long quotations of several paragraphs. It would be
suitable for quoting a few paragraphs from the text of some speech, for instance. LaTEX
treats the given text just like normal text that it has to set into paragraphs, except that
it indents the margins a little.

2.4.16 verse environment

This is provided to facilitate the setting of poetry. When within the verse environment,
we use \newline (or \\) to end a line; and what would normally signify a new paragraph
serves to indicate the start of a new stanza. Let’s have a shot at some cheap poetry.
\begin{verse}
Roses are red\\
Violets are blue\\
I think \TeX\ is great\\
And so will you!

Roses are still red\\

Violets are still blue\\
I’m schizophrenic\\
And so am I.
\end{verse}

will produce the following stunningly-creative “poem”:

24
 Roses are red
Violets are blue
I think TEX is great
And so will you!
Roses are still red
Violets are still blue
I’m schizophrenic
And so am I.

2.4.17 center environment

This environment allows the centring of consecutive lines of text, new lines being indicated
by a \\. If you don’t separate lines with the \\ command then you’ll get a centred
paragraph the width of the page, which won’t look any different to normal. If only one
line is to be centred, then no \\ is necessary.
The {\tt center} environment takes care of the vertical
spacing before and after it, so we don’t need to leave any.
\begin{center}
If we leave no blank line after the\\
{\tt center} environment\\
then the ensuing text will not\\
be regarded as part of a new\\
paragraph, and so will not be indented.\\
\end{center}

In this case we left a blank line after the environment,

so the new text was regarded as starting a new paragraph.

gives the following text

The center environment takes care of the vertical spacing before and after it, so
we don’t need to leave any.

If we leave no blank line after the

center environment
then the ensuing text will not
be regarded as part of a new
paragraph, and so will not be indented.

In this case we left a blank line after the environment, so the new text was
regarded as starting a new paragraph.

2.4.18 flushright and flushleft environments

The flushright environment causes each line to be set with its last character against the
right margin, without trying to stretch the line to the current text width. The flushleft
environment is similar.

25
 We can stop \LaTeX\ from justifying each line to both the
left and the right margins.
\begin{flushright}
The {\tt flushright} environment is\\
used for text with an even right margin\\
and a ragged left margin.
\end{flushright}
\begin{flushleft}
and the {\tt flushleft} environment is\\
used for text with an even left margin\\
and a ragged right margin.
\end{flushleft}

gives the desired display

We can stop LaTEX from justifying each line to both the left and the right margins.

The flushright environment is

used for text with an even right margin
and a ragged left margin.

and the flushleft environment is

used for text with an even left margin
and a ragged right margin.

One must be wary not to lapse into “word-processing” mode when using these environ-
ments. Remember that pressing return at the end of a line in the input file does not serve
to end the current line there, but just to indicate the end of another word. We have to use
the \\ command to end a line.

2.4.19 verbatim environment

We can simulate typed text using the verbatim environment. The \tt (typewriter text)
type style can be used for simulating typed words, but runs into trouble if one of the
characters in the simulated typed text is a specially reserved LaTEX character. For instance,
{\tt type \newline} would not have the desired effect because LaTEX would interpret
the \newline is an instruction to start a new line.
The verbatim environment allows the simulation of multiple typed lines. Everything
within the environment is typeset in typewriter font exactly as it appears in our source
file—obeying spaces and line breaks as in the source file and not recognising the existence
of any special symbols.

26
 \begin{verbatim}
In the verbatim environment we can type anything
we like.
So we don’t need to look out for uses of %, $, & etc,
nor will control sequences like \newline have any
effect.
\end{verbatim}

will produce the simulated input text

In the verbatim environment we can type anything
we like.
So we don’t need to look out for uses of %, $, & etc,
nor will control sequences like \newline have any
effect.

The only thing that cannot be typed in the verbatim environment is the sequence
\end{verbatim}. You might notice that I still managed to simulate that control sequence
above. One can always get what you want in TEX, perhaps with a little creativity.
If we want only to simulate a few typed words, such as when I say to use \newline
to start a new line, then the \verb command is used. This command has a slightly odd
syntax, pressed upon it by the use for which it was intended. It cannot accept an argument,
because we may want to simulate typed text that is enclosed by {braces}. What one does
is to choose any character that is not in the text to be simulated, and use a pair of these
characters as “argument delimiters”. I usually use the @ or " characters, as I rarely have
any other uses for them. Thus
use % to obtain a % sign
is typed as
use \verb"\%" to obtain a \% sign

2.4.20 itemize, enumerate, description environments

LaTEX provides three predefined list-making environment, and a “primitive” list environ-
ment for designing new list environments of your own. We shall just describe the predefined
ones here.
There is delightfully little to learn in order to be able to create lists. The only new
command is \item which indicates the beginning of a new list item (and the end of the
last one if this is not the first item). This command accepts an optional argument (which
means you’d enclose it in square brackets) that can be used to provide an item label. If no
optional argument is given, then LaTEX will provide the item label for you; in an itemize
list it will bullet the items, in an enumerate list it will number the items, and in a list
of descriptions the default is to have no label (which would look a bit odd, so you’re

27
expected to use the optional argument there).
Remember that \item is used to separate list items; it does not accept the list item as
an argument.
\begin{itemize}
\item an item is begun with \verb@\item@
\item if we don’t specify labels, then
\LaTeX\ will bullet the items for us
\item I indent lines after the first in the
input file, but that is just to keep things
readable. As always, \LaTeX\ ignores additional
spaces.

\item a blank line between items is ignored, for

\LaTeX\ is responsible for spacing items.
\item \LaTeX\ is in paragraph-setting mode when
it reads the text of an item, and so will
perform all the usual functions
\end{itemize}

produces the following itemised list:

• an item is begun with \item
• if we don’t specify labels, then LaTEX will bullet the items for us
• I indent lines after the first in the input file, but that is just to keep things readable. As always,
LaTEX ignores additional spaces.
• a blank line between items is ignored, for LaTEX is responsible for spacing items.
• LaTEX is in paragraph-setting mode when it reads the text of an item, and so will perform all the
usual functions

Lists can also be embedded within one another, for they are just environments and
we said that environments have this property. Remember that we must nest them in the
correct order. We demonstrate with the following list, which also shows how to use the
enumerate environment.

28
 \noindent I still have to do the following things:
\begin{enumerate}
\item Sort out LAN accounts for people on the course
\begin{itemize}
\item Have new accounts created for those not already
registered on the LAN
\item Make sure all users have a personal directory
on the data drive
\item Give read and scan rights to users in the \TeX\
directories
\item Add users to the appropriate LAN print queues
\end{itemize}
\item Have a \TeX\ batch file added to a directory that
is on a public search path
\item Finish typing these course notes and proof-read them
\item Photocopy and bind the finished notes
\end{enumerate}

will give the following list

I still have to do the following things:
1. Sort out LAN accounts for people on the course
• Have new accounts created for those not already registered on the LAN
• Make sure all users have a personal directory on the data drive
• Give read and scan rights to users in the TEX directories
• Add users to the appropriate LAN print queues
2. Have a TEX batch file added to a directory that is on a public search path
3. Finish typing these course notes and proof-read them
4. Photocopy and bind the finished notes

See how I lay the source file out in a readable fashion. This is to assist myself, not
LaTEX.
The description environment is, unsurprisingly, for making lists of descriptions.
\begin{description}
\item[\tt itemize] an environment for setting itemised lists.
\item[\tt enumerate] an environment for setting numbered lists.
\item[\tt description] an environment for listing descriptions.
\end{description}

will typeset the following descriptions:

itemize an environment for setting itemised lists.
enumerate an environment for setting numbered lists.
description an environment for listing descriptions.

29
 Note that the scope of the \tt commands used in the item labels was restricted to the
labels.

2.4.21 tabbing environment

This environment simulates tabbing on typewriters. There one chose the tab stops in
advance (analysing the material to be typed for the longest item in each column) and
typed entries consecutively, hitting the tab key to move to the next tab stop and return to
move to the next line.
In the tabbing environment, we proceed similarly. We look for the worst-case line
(that which will determine the desired tab stops) and use it to set the tabs by inserting \=
control symbols at the points where we want tab stops. We then discard that line using
\kill, since the worst-case line might not be the first line in the material we have to type.
We then type each line, using \> to move to the next tab stop and \\ to end a line.
\begin{tabbing}
Cheddar cheese \= Recommended \= \$2.00 \kill
Green Ham \> Recommended \> \$2.00
Eggs \> 1 a week \> \$1.50
Cheddar cheese \> Hmmm \> \$0.80
Yak cheese \> Avoid \> \$0.05
\end{tabbing}

gives the following uniformly-tabbed table

Green Ham Recommended $2.00

Eggs 1 a week $1.50
Cheddar cheese Hmmm $0.80
Yak cheese Avoid $0.05

In the format line I chose the longest entry ¿from each of the prospective columns. I
lined some of the \> commands up in the source just to keep things readable.
Remember that excess spaces are ignored. LaTEX sets the \killed line normally and
sees where the tab stops requested will be needed in the typeset text. Note also that
commands given within the tabbing environment are local to the current item.
Actually, we use the above approach in the case that we require uniformly tabbed
columns. The format line is not compulsory, and we can define tab stops dynamically. See
if you can make sense of the following.
\begin{tabbing}
Entry in position 1,1 \= Entry 1,2 \= Entry 1,3\\
Entry in position 2,1 \> Entry 2,2 \> Entry 2,3\\
Entry 3,1 \= Entry 3,2 \> Entry 3,3\\
Entry 4,1 \> Entry 4,2 \> Entry 4,3
\end{tabbing}

30
which produces

Entry in position 1,1 Entry 1,2 Entry 1,3

Entry in position 2,1 Entry 2,2 Entry 2,3
Entry 3,1 Entry 3,2 Entry 3,3
Entry 4,1 Entry 4,2 Entry 4,3

The are additional commands that can be used within the tabbing environment to
achieve special effects, but we won’t be discussing them here.

2.4.22 tabular environment

The tabular environment is used to produce tables of items, particularly when the table
is predominantly rectangular and when line drawing is required. LaTEX will make most
decisions for us; for instance it will align everything for us without having to be told which
are the longest entries in each column.
This environment is the first of many that use the TEX “tabbing character” &. This
character is used to separate consecutive entries in a row of a table, array, etc. The end of
a row is indicated in the usual manner, by using \\. In this way the individual cells of the
table, or array, are clearly described to LaTEX, and it can analyse them to make typesetting
decisions. Commands issued within a cell so defined are, again, local to that cell.
The tabular environment is also our first example of an environment with arguments.
The arguments are given, in braces as usual, just after the closing brace after the envi-
ronments name. In the case of tabular there is a single mandatory argument giving the
justification of the entries in each column: l for left justified, r for right justified, and c
for centred. There must be an entry for each column of the table, and there is no default.
Let’s start with a simple table.
\begin{tabular}{llrrl}
\bf Student name & \bf Number & \bf Test 1 & \bf Test 2 & \bf Comment\\
F. Basset & 865432 & 78 & 85 & Pleasing\\
H. Hosepipe & 829134 & 5 & 10 & Improving\\
I.N. Middle & 853931 & 48 & 47 & Can make it
\end{tabular}

will produce the following no-frills table

Student name Number Test 1 Test 2 Comment
F. Basset 865432 78 85 Pleasing
H. Hosepipe 829134 5 10 Improving
I.N. Middle 853931 48 47 Can make it
Note that a \\ was not necessary at the end of the last row. Also note that, once
again, the alignment of the & characters was for human readability. It is conventional to
set columns of numbers with right justification. The \bf directives apply only the entries
in which they are given.

31
 A | typed in the tabular environment’s argument causes a vertical line to be drawn
at the indicated position and extending for the height of the entire table. An \hline given
in the environment draws a horizontal line extending the width of the table to be drawn at
the vertical position at which the command is given. A \cline{i-j} draws a line spanning
columns i to j, at the vertical position at which the command is given. A repeated line-
drawing command causes a double line to be drawn. We illustrate line drawing in tables by
putting some lines into our first table. We will type this example in a somewhat expanded
form, trying to make it clear why the lines appear where they do.
\begin{tabular}{|l|l|r|r|l|}
\hline
\bf Student name & \bf Number & \bf Test 1 & \bf Test 2 & \bf Comment\\
\hline
F. Basset & 865432 & 78 & 85 & Pleasing\\
\hline
H. Hosepipe & 829134 & 5 & 10 & Improving\\
\hline
I.N. Middle & 853931 & 48 & 47 & Can make it\\
\hline
\end{tabular}

which will give

Student name Number Test 1 Test 2 Comment
F. Basset 865432 78 85 Pleasing
H. Hosepipe 829134 5 10 Improving
I.N. Middle 853931 48 47 Can make it
That way of laying out the source file makes it clear where the lines will go. As we (by
now) well know, the returns that we pressed after the \\s in typing this table might as
well have been spaces as far as LaTEX is concerned. Thus it is common to have the \hline
commands following the \\s on the input lines. We will do this in future examples.
The \multicolumn column can be used to overrule the overall format of the table for
a few columns. The syntax of this command is
\multicolumn {n}{pos}{item}
where n is the number of columns of the original format that item is to span, and pos
specifies the justification of the new argument.
\begin{tabular}{||l|c|c|c||} \hline
\multicolumn{4}{|c|}{\LaTeX\ size changing commands}\\ \hline
Style option & 10pt (default) & \tt 11pt & \tt 12pt\\ \hline
\tt\bs footnotesize & 8pt & 9pt & 10pt\\ \hline
\tt\bs small & 9pt & 10pt & 11pt\\ \hline
\tt\bs large & 12pt & 12pt & 14pt\\ \hline
\end{tabular}

produces the following table:

32
 LaTEX size changing commands
Style option 10pt (default) 11pt 12pt
\footnotesize 8pt 9pt 10pt
\small 9pt 10pt 11pt
\large 12pt 12pt 14pt

2.4.23 figure and table environments

Figures (diagrams, pictures, etc.) and tables (perhaps created with the tabular environ-
ment) cannot be split across pages. So LaTEX provides a mechanism for “floating” them to
a nearby place where there is room for them. This may mean that your figure or table may
appear a little later in the document than its declaration in the source file might suggest.
You can suggest to LaTEX that it try to place the figure or table at the present position if
there is room or, failing that, at the top or bottom of the present or following page. You
can also ask for it to be presented by itself on a “page of floats”.
You suggest these options to LaTEX through an optional argument to the environment.
One lists a combination of the letters h, t, b, and p where

h means that the object should be placed here if there is room, so that things will appear
in the same order as in the source file,

t means that the object can be placed at the top of the of a text page, but no earlier than
the present page.

b means that the object can be placed at the bottom of a text page, but no earlier than
the present page.

p means that the object should be set on a page of floats that consists only of tables and
figures.

A combination of these indicates decreasing order of preference. The default is tbp. In

this document I have tended to use htbp.
LaTEX will also number and caption a figure or table for you, and compile a list of
tables and a list of figures. Just include \listoffigures and \listoftables next to
your \tableofcontents command at the beginning of the document. To caption a table
of figure, include \caption{caption text} just before the \end{table} or \end{figure}
command. Here’s a sample source file.
\begin{table}[htbp]
\begin{tabular}{lrll}
...
\end{tabular}
\caption{Mark analysis}
\end{table}

To leave space for a figure that will inserted by some other means at a later date, we
can use the \vspace command:

33
 \begin{figure}[htbp]
\vspace{9.5cm}
\caption{An artists impression}
\end{figure}

Including graphics files prepared with drawing packages is possible, but beyond the
scope of this introduction.

2.4.24 The letter document style

All this and we still don’t know how to prepare a simple letter! Actually, there is very
little to it.
Figure 2.1 shows a sample letter. We declare our own address and signature before
entering the letter environment because we can use repeated letter environments to
prepare multiple letters from the same source file. The address of the intended recipient
of a particular letter is given as an argument to the letter environment.

2.4.25 Common pitfalls; Error messages

By now it should be clear that we have to work quite accurately when preparing a docu-
ment. Typing errors in the running text can be absorbed, but messing up a control sequence
name will halt the compiler with an error message. Before we look at some common errors
and some ways to avoid them, let’s have a look at a sample error message.
You’ll have noticed by now that when you run LaTEX on a source file, the transcript of
the compiler session is written on a log file. When errors have accumulated to the point
that LaTEX is hopelessly confused, it is time to debug your source file. The log file contains
a reference to the line, or lines, of your source file that generated the error together with
a description of the error.
TEX and LaTEX error messages appear frightening at first sight, to say the least. They
are actually very informative, but they can take some getting used to. Mistyped control
sequences cause little pain, but a missing \end{environment} can cause a good deal of
confusion because it has the effect of making LaTEX try to set material into that environment
that was never intended/designed to fit in such a place. Also, ommiting a mandatory
argument can cause great confusion.
Suppose we type \bold instead of \bf in the following line:
this is going to be {\bold very} messy.

This produces the following error message:

34
\documentstyle[12pt]{letter}
\begin{document}
\address{(Underneath) Lion Bridge\\
Midway down Commercial Road\\
Pietermaritzburg\\
3200}

\signature{F. Basset\\
Public nuisance}

\begin{Letter}{Director of Public Parks\\

Pietermaritzburg Municipality\\
Pietermaritzburg}

\opening{Kind Sir/Madam}

I wish to complain about the shocking practice of

fencing off the base of trees. I notice with grave
concern that this has occurred in the park bordering
my stately residence.

This has already caused me great inconvenience

and public embarrassment, as you can imagine it would
for a hound of my social standing. I demand that
you take these obscene obstructions away without
delay.

\closing{Yours anxiously}
\end{letter}
\end{document}

Figure 2.1: A sample letter

35
 ! Undefined control sequence.
1.683 this is going to be {\bold
very} messy.
?

That not so bad! The line beginning with ! tells us that we have tried to used a control
sequence that was not known to LaTEX; the 1.683 tells us that the error occurred on line
683 of the source file; and the error message is split over two lines with the break occurring
at the point where LaTEX detected a problem.
But suppose we try the following
\begin{tabular}{llrrl}
Student name & Number & Test 1 & Test 2 & Comment\\
F. Basset & 865432 & 78 & 85 & Pleasing\\
H. Hosepipe & 829134 & 5 & 10 & Improving\\
I.N. Middle & 853931 & 48 & 47 & Can make it

This shows that H.~Hosepipe’s newfound concentration has...

i.e., we omit to provide the \end{tabular} that delimits the end of the environment.
Not having been told that the environment is supposed to be concluded, LaTEX will try to
set the text of the next paragraph as a table item—and will scream blue murder when it
finds that it doesn’t conform to the syntax demanded.
LaTeX error. See LaTeX manual for explanation.
Type H <return> for immediate help.
! \begin{tabular} ended by \end{document}.
\@latexerr ...diate help.}\errmessage {#1}

\@checkend ...urrenvir \else \@badend {#1}

\fi
\enddocument ->\@checkend {document}
\clearpage \begingroup \if@filesw...
\end #1->\csname end#1\endcsname
\@checkend {#1}\expandafter \endgrou...
l.58 \end{document}

Now that’s informative! Actually it is if we agree to ignore all but the the error
indication line (the one beginning with the !) and the line telling us where LaTEX noticed
that all was not well (the one beginning 1.58 in this case). The rest of the error message
you can regard as being for your local TEX wizard to sort your problem out if you are unable
to after consulting the manuals. Tough as it looks, we can decipher this message straight
away: the error indicator line tell us that a tabular environment was ended incorrectly
(in this case by an \end{document}).
TEX error messages aren’t all that bad once you’ve made enough errors to get used

36
to a few! Most can be avoided through careful preparation of the source file. Typing
accurately and knowledge of the command syntax is a good start, but there are some other
precautions that make good sense:

1. Even if LaTEX is happy with free-form input, try to lay your input file out as regularly
and logically as possible. See our examples of environments for formats to adopt.

2. It is important that all group delimiters be properly matched, i.e., braces and \begin{}. . . \end{}
must come in pairs. A good habit to fall in to is to always type such things in pairs
and then move the cursor back between them and type the intervening material.

3. Don’t forget command arguments when they are mandatory. Always ask yourself
what a particular commands needs from you in order to make the decisions that are
required of it.

4. Remember the 10 characters that are specially reserved for commenting, table item
separation, etc.

5. When we look at mathematical typesetting in the next chapter, we will see that the
same principles apply there.

6. Try to use a text editor that has a TEX mode, or at least one that will match brackets
for you.

2.5 Summary
We have learned pretty much all we need to know in order to prepare non-mathematical
documents. There has been quite a lot of material, all told, but we’re fortunate that the
average document requires only a fraction of what we’ve listed here. Furthermore, we’ll
find that what we’ve learned equips us with a good deal of the framework needed for
mathematical typesetting.
The important thing to extract from this chapter is some feel for what I termed the
“spirit of TEX” at the chapter beginning. I cannot emphasise enough the importance of
getting your mind out of “word processing” mode and into “typesetting” mode. Always
keep uppermost in your mind the task at hand: you are to describe the logical content
of the document to LaTEX, so furnishing it with enough information to perform all the
formatting for you.
Many of the earlier sections of this chapter will become trivially easy to you after
just a little experimentation with LaTEX. The best way to learn the syntax of the more
complicated environments is to use them—try typesetting the examples, for instance. It is
important that you come to terms with the tabular environment, for its syntax is typical
of many of the mathematical constructs that we will use.
If you have not already done so, then now is the time to try preparing some documents
of your own. Try including all the material from this chapter, for that is the best way to

37
remember it all. When the initial lack of familiarity wears off, you’ll find that LaTEX is
really a whole lot friendlier and easier to use than you expected.
We must also recognise that there is a lot more to some of the command than detailed
here. Some accept optional argumnets that were not mentioned, others have more options
than we considered. And even once we have a full description of each command, there is
still much to be learned for there is much that can be achieved through creative use of
some of the environments.

38
Chapter 3

Mathematical typesetting with LaTEX

The last chapter taught us a good deal of what we need to know in order to prepare quite
complicated non-mathematical documents. There are still a number of useful topics that
we have not covered (such as cross-referencing), but we’ll defer discussion of those until
a later chapter. In the present chapter, we’ll learn how LaTEX typesets mathematics. It
should come as no surprise that LaTEX does most of the work for us.

3.1 Introduction
In text-only documents we saw that our task was to describe the logical components of
each sentence, paragraph, section, table, etc. When we tell LaTEX to go into mathematical
mode, we have to describe the logical parts of a formula, matrix, operator, special symbol,
etc. TEX has been taught to recognize a binary operation, a binary relation, a variable, an
operator that expects limits, and so on. We just need to supply the parts that make up
each of these, and TEX will take care of the rest. It will leave appropriate space around
operators, italicise variables, set an operator name in roman type, leave the correct space
after colons, place sub- and superscripts in the correct positions (based on what it is you’re
working with), choose the correct typesizes, . . . the list of things it has been taught is
enormous. When you want to revert to setting normal text again, you tell LaTEX too leave
maths mode and go back into the mode it was in (paragraphing mode).
LaTEX cannot be expected to perform these mode shifts itself, for it is not always clear
just when it is mathematics that has been typed. For example, should an isolated letter
a in the input file be regarded as a word (as in the definite article) or a mathematical
variable (as in the variable a). There are no reliable rules for LaTEX to make such decisions
by, so the begin-math and end-math mode switching is left entirely to you.
The symbol $ is specially reserved1 by LaTEX as the “math shift” symbol. When LaTEX
starts setting a document it is in paragraphing mode, ready to set lines of the input file into
paragraphs. It remains in this mode until it encounters a $ symbol, which shifts LaTEX
into mathematical mode. It now knows to be on the look-out for the components of a
1
See section 2.4.4

39
mathematical expression, rather than for words and paragraphs. It reads everything up to
the next $ sign in this mathemtical mode, and then shifts back to paragraphing mode (i.e.
the mode it was in before we took it in to maths mode).
You must be careful to balance your begin-math and end-math symbols. It is often
a good idea to type two $ symbols and then move back between them and type the
mathematical expression. If the math-shift symbols in a document are not matched, then
LaTEX will become confused because it will be trying to set non-mathematical material as
mathematics.
For those who find having the same symbol for both math-begin and math-end con-
fusing or dangerous, there are two control symbols that perform the same operations: the
control symbol \( is a begin-math instruction, and the control symbol \) is an end-math
instruction. Since it is easy to “lose” a $ sign when typing a long formula, a math environ-
ment is provided for such occasions: you can use \begin{math} and \end{math} as the
math-shift instructions. Of course, you could just decide to use $ and take your chances.
Let’s have a look at some mathematics.
\LaTeX\ is normally in paragraphing mode, where
it expects to find the usual paragraph material. Including
a mathematical expression, like $2x+3y - 4z= -1$, in the
paragraph text is easy. \TeX\ has been taught to recognize
the basic elements of an expression, and typeset them appropriately,
choosing spacing, positioning, fonts, and so on.
Typing the above expression without entering maths
mode produces the incorrect result: 2x+3y - 4z= -1

will produce the following paragraph

LaTEX is normally in paragraphing mode, where it expects to find the usual
paragraph material. Including a mathematical expression, like 2x + 3y − 4z = −1, in
the paragraph text is easy. TEX has been taught to recognize the basic elements of
an expression, and typeset them appropriately, choosing spacing, positioning, fonts,
and so on. Typing the above expression without entering maths mode produces the
incorrect result: 2x+3y - 4z= -1
Notice that LaTEX sets space around the binary relation = and space around the binary
operators + and − on the left hand side of the equation, ignoring the spacing we typed in
the input. It was also able to recognize that the −1 on the right hand side of the equation
was a unary minus—negating the 1 rather than being used to indicate subtraction—and
so did not put space around it. It also italicised the variables x, y, and z. However, it did
not italicise the number 1.
In typing a mathematical expression we must remember to keep the following in mind:

1. All letters that are not part of an argument to some control sequence will be italicised.
Arguments to control sequences will be set according to the definition of the command
used. So typing $f(x)>0 for x > 1$ will produce

f (x) > 0f orx > 1

40
 instead of the expression
f (x) > 0 for x > 1
that we intended. Numerals and punctuation marks are set in normal roman type
but LaTEX will take care of the spacing around punctuation symbols, as in

$f(x,y) \geq 0$

which produces
f (x, y) ≥ 0 .

2. Even a single letter can constitute a formula, as in “the constant a”. To type this
you enter $a$ in your source file. If you do not go in to maths mode to type the
symbol, you’ll get things like “the constant a”.

3. Some symbols have a different meaning when typed in maths mode. Not only do
ordinary letters become variables, but symbols such as - and + are now interpreted
as mathematical symbols. Thus in maths mode - is no longer considered a hyphen,
but as a minus sign.

4. LaTEX ignores all spaces and carriage returns when in maths mode, without exception.
So typing something like the constant$ a$ will produce “the constanta”. You
should have typed the constant $a$. LaTEX is responsible for all spacing when in
maths mode, and (as in paragraphing mode) you have to specially ask to have spacing
changed. Even if LaTEX does ignore all spaces when in maths mode you should (as
always in TEX) still employ spaces to keep your source file readable.
The above means that, at least for most material, a typist need not understand the
mathematics in order to typeset it correctly. And even if one does understand the math-
ematics, LaTEX is there to make sure that you adhere to accepted typesetting conventions
(whether you were aware of their existence or not).
So one could type either
$f(x, y) = 2 (x+ y)y/(5xy - 3 )$

or

$f(x,y) = 2(x+y)y / (5xy-3)$

and you’d still get the correct result

f (x, y) = 2(x + y)y/(5xy − 3) .

There are some places where this can go wrong. For instance, if we wish to speak of
the x-y plane then one has to know that it is an endash that is supposed to be placed
between the x and the y, not a minus sign (as $x-y$ would produce). But typing $x--y$

41
will produce x − −y since both dashes are interpreted as minus signs. To avoid speaking
of the x − y plane or the x − −y plane, we should type it as the $x$--$y$ plane. We are
fortunate that LaTEX can recognise and cope with by far the majority of our mathematical
typesetting needs.
Another thing to look out for is the use of braces in an expression. Typing
${x : f(x)>0}$

will not produce any braces. This is because, as we well know, braces are reserved for
delimiting groups in the input file. Looking back to section 2.4.4, we see how it should be
done:
$ \{ x: f(x)>0 \} $

Math shift commands also behave as scope delimiters, so that commands issued in an
expression cannot affect anything else in a document.

3.2 Displaying a formula

LaTEX considers an expression $ ... $ to be word-like in the sense that it considers it to
be eligible for splitting across lines of a paragraph (but without hyphenation, of course).
LaTEX assigns quite a high penalty to doing this, thus trying to avoid it (remember that
LaTEX tries to minimize the “badness” of a paragraph). When there is no other way, it will
split the expression at a suitable place. But there are some expressions which are just too
long to fit into the running text without looking awkward. These are best “displayed” on
a line by themselves. Also, some expressions are sufficiently important that they should
be made to stand out. These, too, should be displayed on a line of their own.
The mechanism for displaying an expression is the display math mode, which is begun
by typing $$ and ended by typing the same sequence (which again means that we’d better
be sure to type them in pairs). Corresponding to the alternatives \( and \) that we had
for the math shift character $, we may use \[ and \] as the display-math shift sequences.
One can also use the environment
\begin{displaymath} ... \end{displaymath}

which is equivalent to $$ ... $$ and is suitable for use with long displayed expressions.
If you wish LaTEX to number your equations for you you can use the environment

\begin{equation} ... \end{equation}

which is the same as the displaymath environment, except that an equation number will
be generated.
It is poor style to have a displayed expression either begin a paragraph or be a paragraph
by itself. This can be avoided if you agree to never leave a blank line in your input file

42
before a math display.
We will see later how to typeset an expression that is to span multiple lines. For now,
let’s look at an example of displaying an expression:
For each $a$ for which the Lebesgue-set $L_a(f) \neq \emptyset$ we define
$$ % We could have used \begin{displaymath} here
{\cal B}_a(f) = \{ L_{a+r}(f) : r > 0 \},
$$ % and \end{displaymath} here
and these are easily seen to be completely regular.

which produces
For each a for which the Lebesgue-set La (f ) 6= ∅ we define

Ba (f ) = {La+r (f ) : r > 0},

and these are easily seen to be completely regular.

That illustrates how to display an expression, but also shows that we’ve got a lot more
to learn about mathematical typesetting. Before we have a look at how to arrange symbols
all over the show (e.g. the subscripting above) we must learn how to access the multitude
of symbols that are used in mathematical texts.

3.3 Using mathematical symbols

LaTEX puts all the esoteric symbols of mathematics at our fingertips. They are all referenced
by name, with the naming system being perfectly logical and systematic. None of the
control words that access these symbols accepts an argument, but we’ll soon see that some
of them prepare LaTEX for something that might follow. For instance, when you ask for the
P
symbol ‘ ’ LaTEX is warned that any sub- or superscripts that follow should be positioned
appropriately as limits to a summation. In keeping with the TEX spirit, none of this
requires any additional work on your part.
We’ll also see that some of the symbols behave differently depending on where they are
P
used. For instance, when I ask for ni=1 ai within the running text, the limits are places
differently to when I ask for that expression to be displayed:
X
n
ai .
i=1

Again, I typed nothing different here—just asked for display math mode.
It is important to note that almost all of the special maths symbols are unavailable
in ordinary paragraphing mode. If you need to use one there, then use an in-line math
expression $...$.

43
3.3.1 Symbols available from the keyboard
A small percentage of the available symbols can be obtained from just a single key press.
They are + − = < > | / ( ) [ ] and ∗. Note that these must be typed within maths mode
to be interpreted as math symbols.
Of course, all of a–z, A–Z, the numerals 0, 1, 2, . . . , 9 and the punctuation characters ,
; and : are available directly from the keyboard. Alphabetic letters will be assumed to be
variables that are to be italicised, unless told otherwise2 . The numerals receive no special
attention, appearing precisely as in normal paragraphing mode. The punctuation symbols
are still set in standard roman type when read in maths mode, but a little space is left
after them so that expressions like {xi : i = 1, 2, . . . , 10} look like they should. Note that
this means that normal sentence punctuation should not migrate into an expression.

3.3.2 Greek letters

Tables 3.1 and 3.2 show the control sequences that produce the letters of the Greek alpha-
bet. We see that a lowercase Greek letter is simply is accessed by typing the control word
of the same name as the symbol, using all lowercase letters. To obtain an uppercase Greek
letter, simply capitalise the first letter of its name.
Just as $mistake$ produces mistake because the letters are interpreted as variables, so
too will $\tau \epsilon \chi$ produce the incorrectly spaced τ χ if you try to type greek
words like this. TEX can be taught to set Greek, but this is not the way. τχ, incidentally,
is the Greek word for “art” and it is from the initials of the Greek letters constituting this
word that the name TEX was derived. TEX is “the art of typesetting”.

α \alpha β \beta γ \gamma δ \delta

 \epsilon ε \varepsilon ζ \zeta η \eta
θ \theta ϑ \vartheta ι \iota κ \kappa
λ \lambda µ \mu ν \nu ξ \xi
π \pi $ \varpi ρ \rho % \varrho
σ \sigma ς \varsigma τ \tau υ \upsilon
φ \phi ϕ \varphi χ \chi ψ \psi
ω \omega

Table 3.1: Lowercase Greek letters

Γ \Gamma ∆ \Delta Θ \Theta Λ \Lambda

Ξ \Xi Π \Pi Σ \Sigma Υ \Upsilon
Φ \Phi Ψ \Psi Ω \Omega

Table 3.2: Uppercase Greek letters

2
See section 3.4.6

44
 ± \pm ∩ \cap  \diamond ⊕ \oplus
∓ \mp ∪ \cup 4 \bigtriangleup \ominus
× \times ] \uplus 5 \bigtriangledown ⊗ \otimes
÷ \div u \sqcap / \triangleleft  \oslash
∗ \ast t \sqcup . \triangleright  \odot
? \star ∨ \vee ∧ \wedge \bigcirc
† \dagger \ \setminus q \amalg ◦ \circ
‡ \ddagger · \cdot o \wr • \bullet

Table 3.3: Binary Operation Symbols

3.3.3 Calligraphic uppercase letters

The letters A, . . . , Z are available through use of the style changing command \cal. This
command behaves like the other style changing commands \em, \it, etc. so its scope must
be delimited as with them. Thus we can type
... for the filter $\cal F$ we have $\varphi({\cal F}) = \cal G$.

to obtain
for the filter F we have ϕ(F ) = G.
There is no need to tabulate all the calligraphic letters, since they are all obtained by
just a type style changing command. We will just list them so that we can see, for reference
purposes, what they all look like. Here they are:

ABCDEF GHIJ KLMN OPQRST UVWX YZ

3.3.4 Binary operators

LaTEX has been taught to recognise binary operators and set the appropriate space either
side of one—i.e., it sets the first argument followed by a little space, then the operator
followed by the same little space and finally the second argument. Table 3.3 shows the
binary operators that are available via LaTEX control words (recall that the binary operators
+, −, and ∗ can be typed from the keyboard). Here are some examples of their use:

Type To produce
$a+b$ a+b
$(a+b) \ otimes c$ (a + b) ⊗ c
$(a \vee b) \wedge c$ (a ∨ b) ∧ c
$X - (A \cap B) = (X-A) \cup (X-B)$ X − (A ∩ B) = (X − A) ∪ (X − B)

45
3.3.5 Binary relations
LaTEX has been taught to recognize the use of binary relations, too. Table 3.4 shows those
available via LaTEX control words. There are a few that you can obtain directly from the
keyboard: <, >, =, and |.
To negate a symbol you can precede the control word that gives the symbol by a \not.
Some symbols come with ready-made negations, which should be used above the \not’ing
method because the slope of the negating line is just slightly changed to look more pleasing.
Thus \notin should be used above \not\in, and \neq should be used above \not =.
If negating a symbol produces a slash whose horizontal positioning is not to your liking,
then use the math spacing characters described in section 3.4.12 to adjust it.

≤ \leq ≥ \geq ≡ \equiv |= \models

≺ \prec  \succ ∼ \sim ⊥ \perp
 \preceq  \succeq ' \simeq | \mid
 \ll  \gg  \asymp k \parallel
⊂ \subset ⊃ \supset ≈ \approx ./ \bowtie
⊆ \subseteq ⊇ \supseteq ∼
= \cong ./ \Join
< \sqsubset = \sqsupset 6 = \neq ^ \smile
.
v \sqsubseteq w \sqsupseteq = \doteq _ \frown
∈ \in 3 \ni ∝ \propto
` \vdash a \dashv

Table 3.4: Binary relations

3.3.6 Miscellaneous symbols

Table 3.5 shows a number of general-purpose symbols. Remember that these are only
available in maths mode. Note that \imath and \jmath should be used when you need
to accent an i or a j in maths mode3 —you cannot use \i or \j that were available in
paragraphing mode. To get a prime symbol, you can use \prime or you can just type ’
when in maths mode, as in $f’’(x)=x$ which produces f 00 (x) = x.

ℵ \aleph 0 \prime ∀ \forall ∞ \infty

h̄ \hbar ∅ \emptyset ∃ \exists 2 \Box
ı \imath ∇
√ \nabla ¬ \neg 4 \triangle
 \jmath \surd [ \flat 4 \triangle
` \ell > \top \ \natural ♣ \clubsuit
℘ \wp ⊥ \bot ] \sharp ♦ \diamondsuit
< \Re k \| \ \backslash ♥ \heartsuit
= \Im 6 \angle ∂ \partial ♠ \spadesuit
0 \mho

Table 3.5: Miscellaneous symbols

3
See section 3.3.10

46
3.3.7 Arrow symbols
LaTEX has a multitude of arrow symbols, which it will set the correct space around. Note
that vertical arrows can all be used as delimiters—see section 3.3.8. The available symbols
are listed in table 3.6.
← \leftarrow ←− \longleftarrow ↑ \uparrow
⇐ \Leftarrow ⇐= \Longleftarrow ⇑ \Uparrow
→ \rightarrow −→ \longrightarrow ↓ \downarrow
⇒ \Rightarrow =⇒ \Longrightarrow ⇓ \Downarrow
↔ \leftrightarrow ←→ \longleftrightarrow l \updownarrow
⇔ \Leftrightarrow ⇐⇒ \Longleftrightarrow m \Updownarrow
7 → \mapsto 7−→ \longmapsto % \nearrow
←- \hookleftarrow ,→ \hookrightarrow & \searrow
( \leftharpoonup * \rightharpoonup . \swarrow
) \leftharpoondown + \rightharpoondown - \nwarrow
*
) \rightleftharpoons ; \leadsto

Table 3.6: Arrow symbols

3.3.8 Expression delimiters

A pair of delimters often enclose an expression, as in
" # (
a11 a12 x if x < 1
and f (x) = .
a21 a22 x2 if x ≥ 1

LaTEX will scale delimiters to the correct size (determined by what they enclose) for you,
if you ask it to. There are times when you don’t want a delimiter to be scaled, so it is left
up to you to ask for scaling.
To ask that a delimter be scaleable, you precede it by \left or \right according as
it is the left or right member of the pair. Scaled delimiters must be balanced correctly.
It sometimes occurs, as in the right-hand example above, that only one member of a
delimiting pair is to be visible. For this purpose, use the commands \left. and \right.
which will produce no visible delimiter but can be used to correctly balance the delimiters
in an expression. For examples of the use of delimiters, see section 3.4.11 where we learn
about arrays.
Table 3.7 shows the symbols that LaTEX will recognise as delimiters, i.e. symbols that
may follow a \left or a \right. Note that you have to use \left\{ and \right\} in
order to get scaled braces.
Z X
3.3.9 Operators like and
These behave differently when used in display-math mode as compared with in-text math
mode. When used in text, they will appear in their small form and any limits provided

47
 ( ( ) ) ↑ \uparrow
[ ] ] ] ↓ \downarrow
{ \{ } \} l \updownarrow
b \lfloor c \rfloor ⇑ \Uparrow
d \lceil e \rceil ⇓ \Downarrow
h \langle i \rangle m \Updownarrow
/ / \ \backslash
| — k \|

Table 3.7: Delimiters

P
will be set so as to reduce the overall height of the operator, as in N
i=1 fi . When used in
display-math mode, LaTEX will choose to use the larger form and will not try to reduce the
height of the operator, as in
X
N
fi .
i=1

Table 3.8 describes what variable-size symbols are available, showing both the small
(in text) and the large (displayed) form of each. In section 3.4.1 we will learn how to place
limits on these operators.
PX T\ JK
\sum \bigcap \bigodot
QY S[ NO
\prod \bigcup \bigotimes
`a FG LM
\coprod \bigsqcup \bigoplus
Z
R W_ U]
\int \bigvee \biguplus
I
H V^
\oint \bigwedge

Table 3.8: Variable-sized symbols

3.3.10 Accents
The accenting commands that we learned for paragraphing mode do not apply in maths
mode. Consult table 3.9 to see how to accent a symbol in maths mode (all the examples
there accent the symbol u, but they work with any letter). Remember that i and j should
lose their dots when accented, so \imath and \jmath should be used.
There also exist commands that give a “wide hat” or a “wide tilde” to their argument,
\widehat and \widetilde.
û \hat{u} ú \acute{u} ū \bar{u} u̇ \dot{u}
ǔ \check{u} ù \grave{u} ~u \vec{u} ü \ddot{u}
ŭ \breve{u} ũ \tilde{u}

Table 3.9: Math accents

48
3.4 Some common mathematical structures
In this section we shall begin to learn how to manipulate all the symbols listed in section 3.3.
Indeed, by the end of this section we’ll be able to typeset some quite large expressions. In
the section following this we will learn how use various alignment environments that allow
us to prepare material like multi-line expressions and arrays.

3.4.1 Subscripts and superscripts

Specifying a sub- or superscript is as easy as you’d hope—you just give an indication that
you want a sub- or superscript to the last expression and provide the material to be placed
there, and LaTEX will position things correctly. So sub- and superscripting a single symbol,
an operator, or a big array all involve the same input, and LaTEX places the material
according to what the expression is that is being sub- or superscripted:
 2
Y
N a11 a12 a13
 
x2 , Xi ,  a21 a22 a23  .
i=1 a31 a32 a33
To tell LaTEX that you want a single character set as a superscript to the last expression,
you just type a ^ before it. The “last expression” is the preceding group or, if there is no
preceding group, the single character or symbol that the ^ follows:

Type To produce
$x^2$ x2
$a^b$ ab
$Y^X$ YX
$\gamma^2$ γ2
$(A+B)^2$ (A + B)2
h 2 in
x +1
$\left[ \frac{x^2+1}{x^2 + y^2} \right]^n$ x2 +y 2

Subscripts of a single character are equally easy—you just use the underscore character
_ where you used ^ for superscripting:

Type To produce
$x_2$ x2
$x_i$ xi
$\Gamma_1(x)$ Γ1 (x)

Now let’s see how to set a sub- or superscript that consists of more than just one
character. This is no more difficult than before if we remember the following rule: _ and ^
set the group that follows them as a sub- and superscripts to the group that precedes the sub-
and superscript symbols. We see now now that our initial examples worked by considering
a single character to be a group by itself. Here are some examples:

49
 Type To produce
$a^2b^3$ a2 b3
$2^{21}$ 221
$2^21$ 22 1
$a^{x+1}$ ax+1
2
$a^{x^2+1}$ ax +1

$(x+1)^3$ (x + 1)3
$\Gamma_{\alpha\beta\gamma}$ Γαβγ
${}_1A_2$ 1 A2

In the very last example we see a case of setting a subscript to an empty group, which
resulted in a kind of “pre-subscript”. With some imagination this can be put to all sorts
of uses.
In all of the above examples the sub- and superscripts were set to single-character
groups. Nowhere did we group an expression before sub- or superscripting it. Even in
setting the expression (x + 1)3 , the superscript 3 was really only set to the character ). If
we had wanted to group the (x + 1) before setting the superscript, we would have typed
${(x+1)}^3$ which gives (x + 1)3 , with the superscript slightly raised. One has to go
to this trouble because, to most people, something like (xa )b is just as acceptable and as
readable as (xa )b . It also has the advantage of aligning the base lines in expressions such
as
(ab)−2 = [(ab)−1 ]2 = [b−1 a−1 ]2 = b−1 a−1 b−1 a−1
which looks more pleasing than if we use additional grouping to force

(ab)−2 = [(ab)−1 ] = [b−1 a−1 ] = b−1 a−1 b−1 a−1

2 2
,

and the latter has rather more braces in it that require balancing.
Here are some more examples, showing how LaTEX will set things just as we want
without any further work on our part:

Type To produce
z
$x^{y^z}$ xy
2
$2^{(2^2)}$ 2(2 )
2ℵ0
$2^{2^{2^{\aleph_0}}}$ 22
d
$\Gamma^{z_c^d}$ Γz c

We can also make use of empty groups in order to stagger sub- and superscripts to an
expression, as in
$\Gamma_{\alpha\beta}{}^{\gamma}{}_\delta$

50
which will yield
Γαβ γ δ
One can specify the sub- and superscripts to a group in any order, but it is best to be
consistent. The most natural order seems to be to have subscripts first, but you may think
otherwise. It is also a good idea to always include your sub- and superscripts in braces (i.e.
make them a group), whether they consist of just a single character or not. This enhances
readability and also helps avoid the unfortunate case where you believe that a particular
control word gives a single symbol yet it really is defined in terms of several.

3.4.2 Primes
LaTEX provides the control word \prime (0) for priming symbols. Note that it is not
automatically at the superscript height, so that to get f 0 you would have to type
$f^\prime$ .

To make lighter work of this, LaTEX will interpret a right-quote character as a prime if used
in maths mode. Thus we can type
$f’(g(x)) g’(x) h’’(x)$

in order to get
f 0 (g(x))g 0 (x)h00 (x) .

3.4.3 Fractions
LaTEX provides the \frac command that accepts two arguments: the numerator and the
denominator (in that order). Before we look at examples of its use, let us just note that
many simple in-text fractions are often better written in the form num/den, as with 3/8
which can be typed as $3/8$. This is also often the better form for a fraction that occurs
within some expression.

Type To produce
x+1
$\frac{x+1}{x+2}$
x+2
1
$\frac{1}{x^2+1}$ 2
x +1
1 + x2
$\frac{1+x^2}{x^2+y^2} + x^2 y$ + x2 y
x2 + y 2
1
$\frac{1}{1 + \frac{x}{2}}$
1 + x2
1
$\frac{1}{1+x/2}$
1 + x/2

51
3.4.4 Roots
The \sqrt command accepts two arguments. The first, and optional, argument specifies
what order of root you desire if it is anything other than the square root. The second, and
mandatory, argument specifies the expression that the root sign should enclose:

Type To produce
√
$\sqrt{a+b}$ a+b
√
5
$\sqrt[5]{a+b}$ a+b
r
n 1+x
$\sqrt[n]{\frac{1+x}{1+x^2}}$ 2
√1 + x
x+1
$\frac{\sqrt{x+1}} {\sqrt[3]{x^3+1}}$ √3
x3 + 1

3.4.5 Ellipsis
Simply typing three periods in a row will not give the correct spacing of the periods if it is
an ellipsis that is desired. So LaTEX provides the commands \ldots and \cdots. Centered
ellipsis should be used between symbols like +, −, ∗, ×, and =. Here are some examples:

Type To produce
$a_1+ \cdots + a_n$ a1 + · · · + a n
$x_1 \times x_2 \times \cdots \times x_n$ x1 × x2 × · · · × xn
$v_1 = v_2 = \cdots = v_n = 0$ v1 = v2 = · · · = vn = 0
$f(x_1,\ldots,x_n) = 0$ f (x1 , . . . , xn ) = 0

3.4.6 Text within an expression

One can use the \mbox command to insert normal text into an expression. This command
forces LaTEX temporarily out of maths mode, so that its argument will be treated as normal
text. It’s use is simple, but we must be wary on one count: remember that LaTEX ignores
all space characters when in maths mode; so to surround words in an expression that were
placed by an \mbox command by space you must include the space in the \mbox argument.

Type To produce
$f_i(x) \leq 0 \mbox{ for } x \in I$ fi (x) ≤ 0 for x ∈ I
$\Gamma(n)=(n-1)! \mbox{ when $n$ is an integer}$ Γ(n) = (n − 1)! when n is an integer

In section 3.4.12 we’ll learn of some special spacing commands that can be used in math
mode. These are often very useful in positioning text within an expression, enhancing
readability and logical layout.

52
 \arccos \cos \csc \exp \ker \limsup \min \sinh
\arcsin \cosh \deg \gcd \lg \ln \Pr \sup
\arctan \cot \det \hom \lim \log \sec \tan
\arg \coth \dim \inf \liminf \max \sin \tanh

Table 3.10: Log-like functions

3.4.7 Log-like functions

There are a number of function names and operation symbols that should be set in normal
(roman) type in an expression, such as in
f (θ) = sin θ + log(θ + 1) − sinh(θ2 + 1)
and
sin h
=1 .
lim
h→0 h
We know that simply typing $log\theta$ would produce the incorrect result
logθ
and that using $\mbox{log}\theta$ would leave us having to insert a little extra space
between the log and the θ
logθ .
So LaTEX provides a collection of “log-like functions” defined as control sequences. Ta-
ble 3.10 shows those that are available. Here are some examples of their use:

Type To produce
$f(x)=\sin x + \log(x^2)$ f (x) = sin x + log(x2 )
$\delta = \min \{ \delta_1, \delta_2 \}$ δ = min{δ1 , δ2 }
$\chi(X) = \sup_{x\in X} \chi(x)$ χ(X) = supx∈X χ(x)
$\lim_{n \rightarrow \infty} S_n = \gamma$ limn→∞ Sn = γ

Notice how LaTEX does more than just set an operation like sup in roman type. It also
knew where a subscript to that operator should go.

3.4.8 Over- and Underlining and bracing

The \underline command will place an unbroken line under its argument, and the \overline
command will place an unbroken line over its argument. These two commands can also be
used in normal paragraphing mode (but be careful: LaTEX will not break the line within
an under- or overlined phrase, so don’t go operating on large phrases).
You can place horizontal braces above or below an expression by making that expression
the argument of \overbrace or \underbrace. You can place a label on an overbrace
(resp. underbrace) by superscripting (resp. subscripting the group defined by the bracing
command.

53
 Type To produce
$\overline{a+bi} = a- bi$ a + bi = a − bi
$\overline{\overline{a+bi}} = a+bi$ a + bi = a + bi

And some examples of horizontal bracing:

$A^n=\overbrace{A \times A \times \ldots \times A}^{\mbox{$n$ terms}}$

$\forall x \underbrace{\exists y (y \succ x)}_{\mbox{scope of $\forall$}}$

will produce
n terms
z }| {
A = A×A ×...× A
n

and
∀x ∃y(y  x)
| {z }
scope of ∀

3.4.9 Stacking symbols

LaTEX allows you to set one symbol above another through the \stackrel command. This
command accepts two arguments, and sets the first centrally above the second.

Type To produce
f∗
$X \stackrel{f^*}{\rightarrow}Y$ X →Y
4
$f(x) \stackrel{\triangle}{=} x^2 + 1$ f (x) = x2 + 1

3.4.10 Operators; Sums, Integrals, etc.

Each of the operation symbols in table 3.8 can occur with limits. They are specified as sub-
and superscripts to the operator, and LaTEX will position them appropriately. In an in-text
formula they will appear in more-or-less the usual scripting positions; but in a displayed
formula they will be set below and above the symbol (which will also be a little larger).
The following should give you an idea of how to use them:

Type To produce
PN
$\sum_{i=1}^{N} a_i$ ai
R b i=1
$\int_a^b f$ f
Ha
$\oint_{\cal C}f(x)\,dx$ f (x) dx
QC
$\prod_{\alpha \in A} X_\alpha$ α∈A Xα
P
$\lim_{N\rightarrow\infty}\sum_{i=1}^{N}f(x_i)\Delta x_i$ limN →∞ N i=1 f (xi )∆xi

54
 We’ll have more to say about the use of \, in section 3.4.12. Let’s have a look at each
of those expressions when displayed:

X
N Z b I Y X
N
ai , f , f (x) dx , Xα , lim f (xi )∆xi
a C N →∞
i=1 α∈A i=1

3.4.11 Arrays
The array environment is provided for typesetting arrays and array-like material. It
accepts two arguments, one optional and one mandatory. The optional argument specifies
the vertical alignment of the array—use t, b, or c to align the top, bottom, or centre of the
array with the centreline of the line it occurs on (the default being c). The second argument
is as for the tabular environment: a series of l, r, and c’s that specify the number of
columns and the justification of these columns. The body of the array environment uses
the same syntax as the tabular environment to specify the individual entries of the array.
For instance the input
... let $A = \begin{array}{rrr}
12 & 3 & 4\\
-2 & 1 & 0\\
3 & 7 & 9
\end{array}$ ...

will produce the output

 
12 3 4
let A =  −2 1 0 
3 7 9
Note that we had to choose and supply the enclosing brackets ourselves (they are not
placed for us so that we can use the array environment for array-like material; also, we
get to choose what type of brackets we want this way). As in the tabular environment,
the scope of a command given inside a matrix entry is restricted to that entry.
We can use ellipsis within arrays as in the following example:
\det A = \left| \begin{array}{cccc}
a_{11} & a_{12} & \cdots & a_{1n}\\
a_{21} & a_{22} & \cdots & a_{2n}\\
\vdots & \vdots & \ddots & \vdots\\
a_{m1} & a_{m2} & \cdots & a_{mn}
\end{array} \right|

which produces
 
 a11 a12 ··· a1n 
 
 a21 a22 ··· a2n 
 
det A =  . .. .. .. 
 .. . . . 
 
 am1 am2 · · · amn 

55
 The array environment is often used to typeset material that is not, strictly speaking,
an array:
f(x) = \left\{ \begin{array}{ll}
x & \mbox{for $x<1$}\\
x^2 & \mbox{for $x \geq 1$}
\end{array} \right.

which will yield


x for x < 1
f (x) =
x2 for x ≥ 1

3.4.12 Changes to spacing

Sometimes LaTEX needs a little help in spacing an expression, or perhaps you think that
the default spacing needs adjusting. For these purposes we have the following spacing
commands:
\, thin space \: medium space
\! negative thin space \; thick space
\quad a quad of space \qquad two quads of space
The spacing commands \,, \quad, and \qquad can be used in paragraphing mode, too.
Here are some examples of these spacing commands used to make subtle modifications to
some expressions.

Type To produce
√
$\sqrt{2} \, x$ 2x
Rb
$\int_a^b f(x)\,dx$ a
f (x) dx
$\Gamma_{\!2}$ Γ2
R bR d
$\int_a^b \! \int_c^d f(x,y)\,dx\,dy$ a c f (x, y) dx dy
$x / \! \sin x$ x/sin x
√
$\sqrt{\,\sin x}$ sin x

3.5 Alignment
Recall that the $equation$ environment can be used to display and automatically number
a single- line equation4 . The eqnarray environment is used for displaying and automati-
cally numbering either a single expression that spreads over several lines or multiple expres-
sions, while taking care of alignment for us. The syntax is similar to that of the tabular
and array environments, except that no argument is necessary to declare the number and
4
See section 3.2

56
justification of columns. The eqnarray* environment does this without numbering any
equations.
\begin{eqnarray}
(a+b)(a+b) & = & a^2 + 2ab + b^2\\
(a+b)(a-b) & = & a^2 - b^2
\end{eqnarray}

will give

(a + b)(a + b) = a2 + 2ab + b2 (3.1)

(a + b)(a − b) = a2 − b2 (3.2)

See how we identify the columns so as to line the = signs up. We can also leave entries
empty, to obtain effect like the following:
\begin{eqnarray*}
\frac{d}{dx} \sin x & = & \lim_{h\rightarrow0}\frac{\sin(x+h)-\sin x}{h}\\
& = & \lim_{h\rightarrow0}\frac{\sin x\cos h + \cos x\sinh - \sin x}{h}\\
& = & \lim_{h\rightarrow0}\frac{\sin x(\cos h-1)}{h} + \cos x\frac{\sin h}{h}\\
& = & \cos x
\end{eqnarray*}

which produces

d sin(x + h) − sin x
sin x = lim
dx h→0 h
sin x cos h + cos x sinh − sin x
= lim
h→0
( h )
sin x(cos h − 1) sin h
= lim + cos x
h→0 h h
= cos x

3.6 Theorems, Propositions, Lemmas, . . .

Suppose you document contains four kinds of theorem-like structures: “theorems”, “propo-
sitions”, “conjectures”, and “wild guesses”. Then near the beginning of the document you
should have something like the following:
\newtheorem{thm}{Theorem}
\newtheorem{prop}{Proposition}
\newtheorem{conjec}{Conjecture}
\newtheorem{wildshot}{Hypothesis} % make it sound good!

57
 The first argument to \newtheorem defines a new theorem-like environment name of
your own choosing. The second argument contains the text that you want inserted when
your theorem is proclaimed:
\begin{thm} $X$ is normal if, and only if, each pair of disjoint
closed sets in $X$ is completely separated.
\end{thm}

\begin{wildshot} % remember, we chose the name ’wildshot’

The property of Moore extends to all objects of the class $\Sigma$.
\end{wildshot}

which will produce the following:

Theorem 1 X is normal if, and only if, each pair of disjoint closed sets in X is completely
separated.

Hypothesis 1 The property of Moore extends to all objects of the class Σ.

Notice that LaTEX italicises the theorem statement, and that you still have to shift in
to maths mode when you want to set symbols and expression. Typically, it is the style
file that determines what a theorem will appear like—so don’t go changing this if you
are preparing for submission for publication (because the journal staff want to substitute
their production style for your document style choice, and not be over-ridden by other
commands).

3.7 Where to from here?

We have covered a good deal of LaTEX’s mathematical abilities, albeit rather superficial
coverage here and there. There is much that has been impressive, but there is clearly a
lot more to TEXnical typesetting than we have covered here—it is not difficult to think of
an expression that we don’t yet know how to typeset. Also, there are places where LaTEX
is a little weak and it leaves us to do somewhat more work than the spirit of TEX would
suggest.
Of course, we cannot criticise LaTEX until we know its full capability. So the first place
to go from here is the LaTEX User’s Guide & Reference Manual. Particularly, the command
reference guide in Appendix C of that book is an invaluable source of LaTEX information
that few can afford to do without. With good knowledge of the LaTEX environments and
their options (and we’ve left out many here) one can accomplish a good deal of most
typesetting problems. A little imagination (say putting an environment to a slightly non-
standard use) can often solve more difficult problems. Lastly, of course, much of raw
TEX still sits underneath LaTEX and so it is true to say that you can do anything with
LaTEX—but you may need some divine inspiration from time to time (ask your local TEX
guru).

58
 In the next chapter we will look, very briefly, at a number of LaTEX commands that we
have not yet considered. Nothing exciting on the mathematical front, but there is some
other important material (e.g. cross-referencing and page-sizing). For now, let’s look at
the “way forward” with respect to mathematical typesetting.

3.8 AMS-LaTEX
Back in the introduction we said that AMS-LaTEX was just a big macro package, the result
of a marriage of LaTEX and AMS-TEX designed to endow the powerful general-purpose
LaTEX package with the mathematical prowess of AMS-TEX without compromising the
LaTEX syntax. Most of that is true, except that AMS-LaTEX is really just a document-
style option (like 12pt) that can be used within a “tweaked LaTEX”. The most visible part
of this tweak is the new font selection scheme of Mittelbach and Schöpf, discussed more
fully in the next chapter. Almost every LaTEX command and environment survived the
transition to AMS-LaTEX, the exceptions being those that were considered redundant or
under-used (space is at a premium with such a big package). The tweaked LaTEX package
is therefore able to deal with practically every existing LaTEX document, giving just a few
(often pleasant) surprises.
With the amstex style option, one can just start a LaTEX document with

\documentstyle[amstex]{article} % or report, book, etc

to gain access to the TEXnical excellence of the AMS technical staff. It is not necessary to
have read The Joy of TEX (the AMS-TEX reference guide) to be able to use the amstex
option, for AMS-LaTEX comes with its own reference guide. Even so, The Joy of TeX is
still highly recommended reading. The syntax of AMS-TEX has been changed to that of
LaTEX, but one can perform that transformation as you read “Joy” and still learn much of
the art of technical typesetting. Nowhere else will you find so comprehensive a coverage
of the conventions and pitfalls of mathematical typesetting. In addition, Joy lists all the
extra symbols that are available through the amstex option (if you thought LaTEX had
a fair selection of esoteric symbols, just wait ’til you see those!) and provides in-depth
accounts where the AMS-LaTEX documentation is brief.
Just as AMS-TEX comes with the AMS preprint style (amsppt), AMS-LaTEX comes
with a specialist style file for preparation of articles with AMS-LaTEX for submission to
journals: amsart. The AMS-LaTEX User’s Guide is quite short and very terse in its
explanations (assuming you to be competent in LaTEX), but is supplemented by a large
body of examples and a comprehensive sample article that is a showcase of the abilities of
the amsart style. You must read both these documents to really learn AMS-LaTEX. The
AMS also distributes a guide to authors who wish to submit using AMS-LaTEX, and this
is a must-read once you are familiar with some of AMS-LaTEX.

59