citationcommands.tex

%\chapter{Citation Commands}
\chapter{标注命令}\label{ch:citationcommands}

The heart of any citation package, from the user’s point of view, is
the citation commands. 
\marginpar{\footnotesize 从用户的角度看，引用宏包的核心是标注(引用)命令。
}
These are what you use most. \biblatex\ offers a
number of different commands. They follow a consistent pattern; so
once you've mastered one it’s quite easy to grasp the others. Their
exact output depends on the citation style you are using.

\index{citation!general form}
Let's start not with the commands themselves, but by looking at some
citations. Consider the following:

\begin{quote}
See also [1]

(Jones 1990, p. 24)

cf [JJ90, ch. 3]

see Jones, \emph{Combinatorial Aesthetics} (London: Pubco, 1990)

ibid. 202
\end{quote}

There's a common pattern behind the obvious differences. Each citation
consists of up to three parts: an (optional) `signal' that introduces
it; the citation itself, which directly or indirectly identifies a
source; and an (optional) `pinpoint' which locates a particular part
of the cited work.

\begin{margintable}
\begin{tabular}{lll}
\toprule
\textsf{signal} & \textsf{citation} & \textsf{pinpoint} \\
\midrule
See also &   [1] \\
         &   Jones 1990   & p. 24 \\
cf.      &   JJ90         & ch. 3 \\
see      &   Jones \ldots\    \\
         &   ibid.        &  202 \\
\bottomrule
\end{tabular}
\vspace{3pt}
\caption{The structure of citations}
\end{margintable}

\index{citation!general pattern}\index{prenote}\index{postnote}
The citation commands in \biblatex\ follow the same pattern: they have
one mandatory argument, which specifies the `key' of the source cited,
and two optional arguments. The primary optional argument is called
the \texttt{postnote} and gives the `pinpoint', and the secondary optional
argument is called the \texttt{prenote} and specifies the `signal'.

\begin{table}
\begin{tabular}{llll}
\toprule

                   & \textsf{prenote}  & \textsf{postnote} &  \textsf{key} \\
\midrule
\cs{cite[See also][]\{jones\}}   & See also    &        & jones \\
\cs{cite[24]\{jones\}}           &             & p. 24  & jones \\
\cs{cite[cf.][ch. 3]\{jones\}}   & cf.         & ch. 3  & jones \\
\cs{cite[see][]\{jones\}}        & see         &        & jones \\
\bottomrule
\end{tabular}
\vspace{6pt}
\caption{Citation command with prenotes and postnotes}
\end{table}


Notice that if there is only one optional argument it is assumed to be
the postnote; but if there are two, then the first optional argument
is the prenote. 
\marginpar{\footnotesize 标注命令通常有一个必选参数和两个可选参数，给出一个可选参数时作为postnote。
}
If we want a prenote but no postnote, as in the first
and last examples, we have to leave the second argument blank. But if
we want a postnote and no prenote, we just use one optional
argument. All the citation commands follow this format.

%\section{The Basic Five}
\section{五个基本命令}
\indexstart{citation!commands!general}

There are five basic citation commands: \cs{cite}, \cs{footcite},
\cs{parencite}, \cs{autocite} and \cs{textcite}. Exactly what each one
does depends on the citation style. But the general idea of the
circumstances in which you would use each is more or less the same.

\paragraph{\textbackslash cite.}\csindex{cite} The \cs{cite} command is “basic”. It
simply prints the particular style’s idea of a simple citation. So,
for instance, with a numeric style, \cs{cite} prints [1], and with a
verbose style it prints full bibliographical details (or ibid., or
supra n. \ldots, if that would be correct).


\paragraph{\textbackslash footcite and \textbackslash parencite.}\csindex{parencite} The
\cs{footcite} command tries to put its citation in a footnote. So
\cs{footcite\{jones\}} is means the same thing as \cs{footnote\{\textbackslash
  cite\{jones\}.\}} --- syntactic sugar. The \cs{parencite} command
does a similar thing, but it wraps its citation in parentheses, so
that it is equivalent to \texttt{(\cs{cite\{jones\}})}.


\paragraph{\textbackslash autocite.}\csindex{autocite} This is all very well. But what
if you’re not sure whether you should use \cs{cite} or \cs{footcite}
or \cs{parencite}. Suppose you write a paper, expecting to use
numerical citations; but then you decide to use verbose citations
instead. Now you have to go through and convert all your \cs{cite}s to
\cs{footcite}s. The \cs{autocite} command is intended to avoid
that. It will use whatever citation command is most appropriate to the
style selected. So if you use \cs{autocite}, you can (try to) change
all your citations from in-text citations to, say, footnotes simply by
changing the style.

It does another thing that is, as the kids say, cool. It will shift
punctuation. It needs to be able to do this because where a citation
appears relative to punctuation may vary. In English and American
practice, for instance, footnotes appear after punctuation. So whereas
a correct numerical citation would be `blah [1].' a correct verbose one
would be `blah.$^1$' With `\cs{autocite\{jones\}}.' \biblatex\ will move
punctuation from after the citation command to put the footnote marker
in the right place.

There are, however, limitations. The \cs{autocite} command can cope
well with citations at the end of sentences; but if citations are
entangled in the text too much, the result will not be satisfactory. A
sentence like
\begin{quote}
  Jones (1990) and Smith (1991) `independently concluded' (Doe 2002,
  p.~19) that foos should not be permitted to bar.
\end{quote}
will probably require some manual intervention.

\paragraph{\textbackslash textcite.}\csindex{textcite} Finally there is
\cs{textcite}. It’s not uncommon to have sentences like
\begin{quote}
   \ldots\ as Jones (1990) remarks \ldots
\end{quote}
where the citation is `woven' into the text of the sentence.

In numerical and verbose styles this can be done manually:
\begin{quote}
   \verb|as Jones \cite{jones} remarks| \gives\ as Jones [1] remarks
\end{quote}
But that's a bit wordy, and in author/year styles, this trick won't work:
\begin{quote}
   \verb|as Jones \parencite{jones} remarks| \gives\ as Jones (Jones 1990) remarks
\end{quote}
\cs{textcite} is designed to deal with this. It offers a citation in a form appropriate for running text.
\begin{quote}
   \verb|as \textcite{jones} remarks| \gives\ as Jones (1990) remarks
\end{quote}

Although mainly intended for author/year styles, \cs{textcite} works
rationally in any style; in verbose styles, the proper form involves
printing something in the text (such as `Jones') and adding a
footnote; if necessary punctuation will be moved for this purpose, so
that
\begin{Verbatim}
... as \textcite[11]{Jones}, presciently, remarked ...
\end{Verbatim}
will produce \marginnote{\tikz[overlay]{\draw (-0.5em,-0.5em) -|
    ([shift={(0.5ex,2.3ex)}] pic cs:comma)}The comma is moved so that
  it appears before the footnote.}
\begin{quote}
  \dots as
  Jones\colorbox{red!30}{\tikzmark{comma},\textsuperscript{1}}
  presciently, remarked
\end{quote}

\subsection{Summary}

Table \ref{allcitations} shows how the various citation commands
operate with the various different standard styles.

\begin{table*}
\scriptsize
\begin{tabular}{llllll}
\toprule
                                         & \cs{cite[9]\{jones\}}     & \cs{footcite[9]\{jones\}}  &
\cs{parencite[9]\{jones\}}               & \cs{autocite[9]\{jones\}} & \texttt{as} \cs{textcite[9]\{jones\}}
says                                                                                                                        \\
\midrule \textsf{numeric}                & [1, p.~9]                 & $^*$1,
  p.~9.                                  & [1, p.~9]                 & [1, p.~9]                  & as Jones [1, p.~9] says \\
\midrule \textsf{alphabetic}             & [Jon94, p.~9]             & $^*$Jon94, p.~9            & [Jon94, p.~9]
                                         & [Jon94, p.~9]             & as Jones [Jon94, p.~9] says                          \\
\midrule \textsf{authoryear}             & Jones 1994, p.~9          & $^*$Jones 1994, p.~9       & (Jones
1994, p.~9)                              & (Jones 1994, p.~9)        & as Jones (1994,
p.~9) says                                                                                                                  \\
\midrule \textsf{authortitle}            & Jones, ``Title'', p.~9    & $^*$Jones, ``Title'',
p.~9                                     & (Jones, ``Title'', p.~9)  & $^*$Jones, ``Title'', p.~9 & as
Jones (``Title'', p.~9) says                                                                                                \\
\midrule \textsf{verbose}                & \parbox{2cm}{Jones. ``Title''. In: \emph{Journal}
  100 (1994), p.~7, p.~9}                & \parbox{2.2cm}{$^*$Jones. ``Title''. In:
  \emph{Journal} 100 (1994), p.~7, p.~9} & \parbox{2cm}{(Jones. ``Title''. In: \emph{Journal}
  100 (1994), p.~7, p.~9)}               & \parbox{2cm}{Jones. ``Title''. In: \emph{Journal}
  100 (1994), p.~7, p.~9}                & \parbox{3cm}{as Jones$^*$ says                                                   \\---\\
  $^*$Jones. ``Title''. In: \emph{Journal}
  100 (1994), p.~7, p.~9.}                                                                                                  \\
\bottomrule
\end{tabular}
\vspace{5pt}
\caption{Common citation commands in standard styles\label{allcitations}}
\end{table*}
\indexstop{citation!commands!general}

%\section{Capital Forms}
\section{五个基本命令的首字母大写形式}

\indexstart{citation!commands!capital forms} Each of \cs{cite},
\cs{parencite} and \cs{autocite} have `capital' forms (\cs{Cite}, etc)
which are appropriate for use where a capital would be required. You
might wonder why, since most citations already begin with a capital,
or with a number. The only time capitalisation might be needed is if
the citation begins with something like `von', and you want it
capitalised, or if the footnote reads `ibid' and you need
`Ibid'. There is no special capital form of \cs{footcite} because it
always assumes that the citation will begin a sentence so that a
capital may be appropriate.  \indexstop{citation!commands!capital
  forms}

%\section{Specific to Numerical Styles}
\section{顺序数字编码样式的注意点}

\index{citation!commands!for numerical styles}\csindex{supercite}
For numerical styles, the command \cs{supercite} prints citations as
superscript numbers. It’s rather unusual to want both regular
full-size labels and superscript labels in the same document; but if
you do you can mix \cs{cite} and \cs{supercite} as you wish. 
\marginpar{\footnotesize 注意 supercite命令忽略prenote 和 postnote。
}
Generally
speaking you want one or the other. For this, the safest course is to
use the option \texttt{autocite=super}, and use \cs{autocite}.

There is a limitation to \cs{supercite}. It doesn't make sense to try
to cram signals and pinpoints into superscript numbers. So
\cs{supercite} will discards prenote and postnote parts
\begin{quote}
\cs{supercite[see][25]\{jones\}} \gives\ $^1$
\end{quote}
If this happens, a warning will be given.

%\section{Using particular fields}
\section{标注部分文献信息的命令}

\indexstart{citation!commands!particular fields}
There are a number of citation commands that are not intended to print
complete citations, but to pull potentially useful snippets out of a
source record, and treat it is a citation. They are mostly, though not
exclusively, useful in author/title and author/year styles, where
citations are often quite tightly woven into the text.

\csindex{citeauthor}\csindex{citetitle}\index{citation!commands!author}\index{author!citing name}\index{citation!commands!title}\index{title!citing title}
%\paragraph{Author or title alone.}
\paragraph{仅标注作者或题名} The commands \cs{citeauthor} and
\cs{Citeauthor} print the author's name (the latter with capitals
forced, even if they would otherwise not be used). The commands
\cs{citetitle} and \cs{citetitle*} print the title: abbreviated if
possible in the regular form, but always in full in the starred form.

\index{citation!commands!year}\csindex{citeyear}\csindex{citedate}\index{date!citing date}\index{year|see{date}}

%\paragraph{Year or date.} 
\paragraph{年份或日期} The\marginnote{{\ttfamily
    \cs{citeauthor\{key\}} published his roman a clef
    (\cs{citeyear*}\{key\}) in \cs{citeyear}\{key\}} \gives\ Locke
  published his roman a clef (1905b) in 1905} commands \cs{citeyear}
and \cs{citeyear*}, and \cs{citedate} and \cs{citedate*}, print year
or date. The difference is that whereas the unstarred versions just
print the calendar year or date, the starred versions print any extra
label that is added to that for the particular work. So if you had,
say, five works by a particular author in one year,
\verb|\citeyear{prolific86c}| would print `1986', whereas
\verb|citeyear{prolific86c}| would print `1986c'.

\index{citation!commands!url}\index{URL!citing}\csindex{citeurl}
%\paragraph{URLs.} 
\paragraph{网址} The \cs{citeurl} command will print any url.

%\paragraph{Others.} 
\paragraph{其它} Apart from these commands, which the standard
styles already provide, it is possible to construct your own `partial'
commands,\manref{\S~3.8.7} using \cs{citename}, \cs{citelist} and
\cs{citefield}. There are three commands because \biblatex\
distinguishes between \verb|name| fields (like \texttt{author} and
\texttt{editor}), list fields (like \texttt{institution} and
\texttt{publisher}) and other fields. The commands are not really
suitable or intended for direct use: they would need to be `wrapped'
in some more user-friendly command for use in a document.
\indexstop{citation!commands!particular fields}

%\section{Entry sets}
\section{条目集}

\index{citation!commands!entry sets}\index{sets of sources cited together|see{entryset}}\index{entryset@\texttt{entryset}}
In some disciplines it is common to have a single numeric citation
refer to a group of sources. This is supported in \biblatex\ by the
entry-set.\manref{\S~3.12.5}

\index{entryset@\texttt{entryset}!bib file in@in \texttt{.bib} file}
There are two ways to define such a set. One way is to define them as
a |key| in the |.bib| file. Thus, for instance, your |.bib| file might
define
\begin{Verbatim}
@set{set1,
  entryset={augustine, cotton}}
\end{Verbatim}
You can then use |\cite{set1}| to get output as in figure \ref{entryset1}.
\begin{figure}
\fbox{\includegraphics{examples/entryset1.pdf}}
\caption{An \texttt{entryset}\label{entryset1}}
\end{figure}

\csindex{defbibentryset}\index{entryset@\texttt{entryset}!using defbibentryset@using \cs{defbibentryset}}
The alternative, and usually more convenient approach, is to define
such sets `on the fly' in each document. This can be done using the
command
\begin{pseudoverb}
  \centering\cs{defbibentryset}\{\angled{set-key}\}\{\angled{key$_{1}$
    \ldots\ key$_{n}$}\}
\end{pseudoverb}
so that
\begin{pseudoverb}
  \centering
\cs{defbibentryset}\{set1\}\{augustine, cotton\}
\end{pseudoverb}
will produce the same output as figure \ref{entryset1}.

%\section{Pinpoints}
\section{标注的一些细节}

\indexstart{citation!postnote}\index{pinpoints|see{citation, postnote}}
It usually makes sense to give pinpoint citations to a particular work
primarily by reference to one type of marker. 
\marginpar{\footnotesize 标注还要注意一些细节，比如页码引导符等。输入数字可以让biblatex自动输出引导符，也可以手动输入引导信息。也可以利用 pagination 域，本地化字符串，nopp，pno，ppno，psq(拉丁)，psqq(拉丁)，postnote 域格式等做更细致的处理。
}
So, for instance, books
are normally cited by page. How such references are styled depend on
the particular scheme. Some schemes want `p. 23', others want `p 23',
others just want `23'. To save effort, \biblatex\ will insert
appropriate default labels. So if you just type \cs{cite[23]\{key\}},
\biblatex\ will produce `p. 23' (or whatever is appropriate to your
style). It works also for multiple pages: \cs{cite[22, 23]\{key\}}
will print `pp. 22, 23' and \cs{cite[22--25]\{key\}} will print
`pp. 22--25', and do forth.


What if you want some different prefix? If the problem is with a
particular citation, you can just enter the prefix manually. For
instance, if you want to cite chapter 3, just type
\cs{cite[ch.\textasciitilde 3]\{key\}}. Since \biblatex\ only adds a
prefix where you have provided a reference which consists only of
numbers, your postnote will be printed as you have entered it.

\newthought{This is neat and tidy.} But, as so often there are special
cases and customisations.

\index{citation!postnote!suppress prefixes}
%\paragraph{No prefixes}
\paragraph{不使用引导符} You don't want prefixes, even when you enter a
purely numerical postnote: you want \cs{cite[22]\{key\}} to produce
`\angled{key}, 22'. There are a number of ways to achieve this,
depending on how generally you want to disable the use of prefixes.
For a single citation, start the postnote with the special command
\cs{nopp}.  \marginnote {{\ttfamily \cs{cite}[\cs{nopp}22]\{key\}}
  \gives\ \angled{key}, 22} This is most appropriate where you have a
single citation that is giving you problems.

To\marginnote{%
  {\ttfamily @book\{key,...\\
    \quad pagination = \{none\}\}} \\%
  \cs{cite[22]\{key\}} \gives\ \angled{key}, 22} stop all additions
for a particular source, set that source’s pagination to ’none’. This
is most appropriate when you generally want to use prefixes, but you
have a particular source where they are not required.

To stop all additions, for every source:
\begin{center}
\verb|\DeclareFieldFormat{postnote}{#1}|
\end{center}
This is most appropriate when you are customising \biblatex\ for a
style which does not use any prefixes in general.

\index{citation!postnote!force prefixes}
%\paragraph{Forcing prefixes}
\paragraph{强制使用引导符} You want to enter a postnote with a
non-numerical character, and still get a prefix added. For instance,
you have a source with pages which have letters, like `1234a'. Again,
there are a number of solutions.

\csindex{pno}\csindex{ppno}
The\marginnote{\cs{cite[p.\textasciitilde 12a]\{key\}} \gives\
  \angled{key}, p.~12a} easiest thing to do in such cases is simply to
type the prefix yourself: \cs{cite[p.\textasciitilde
  1234a]\{key\}}. But if you think it’s worth getting \biblatex\ to do
it for you, you can use \cs{pno}\marginnote{\cs{cite[\textbackslash
    pno 12a]\{key\}} \gives\ \angled{key}, p.~12a} (for a single page)
and \cs{ppno}\marginnote{\cs{cite[\textbackslash ppno
    12a-{}-b]\{key\}} \gives\ \angled{key}, pp.~12a--b} (for multiple
pages). (In theory, the \cs{pno} and \cs{ppno} commands could be
`better' because you could change the style of prefixes and they would
adjust automatically. In the real world, there's probably not much
practical advantage.

\index{citation!postnote!Latin gadgets}\csindex{psq}\csindex{psqq}
%\paragraph{Latin gadgets for multiple pages}
\paragraph{拉丁式引导符} There is one special
common case. Suppose you want `p.~20~ff.' or `p.~20 et seq.'  You
could of course enter the whole postnote manually. But you can also
use the special commands \cs{psq}\marginnote{\cs{cite[20\textbackslash
    psqq]\{key\}} \gives\ \angled{key}, pp.~20 sqq.} and \cs{psqq} to
add the indication. The advantage is that the particular phrase to
refer to subsequent page(s) can then be set consistently. You should
not put any space between these commands and the number:
\cs{cite[20\textbackslash psqq]\{key\}}.

By default, in the standard styles, \cs{psq} gives `sq.' and \cs{psqq}
gives `sqq.' To redefine them, redefine the bibliography strings
\verb|sequens| (for a single following page), and \verb|sequentes|
(for more than one). For instance, to use `et seq.' for both, one
would (assuming we are using American English)
\begin{verbatim}
\DefineBibliographyStrings{american}{%
    sequens = {et seq\adddot},
    sequentes = {et seq\adddot}}
\end{verbatim}
The space before this addition is added by a command called
\cs{sqspace}. Some citation styles don't add any space at this point,
preferring `10f.' to `10 f.' If you don't want any space, redefine
this command:
\begin{center}
\verb|\renewcommand{\sqspace}{}|
\end{center}

%\paragraph{Correcting \biblatex's guesses}
\paragraph{修正 biblatex 的判断错误} \biblatex\ tries to
guess whether to add `p.' or `pp.' by working out whether you have
cited a single page or a range of pages. But occasionally the form of
a citation might confuse \biblatex, so that it adds `p.' when you need
`pp.', or vice versa. In such cases, either type the whole postnote
yourself, or use the \cs{pno} or \cs{ppno} commands, as explained
above.

\index{citation!postnote!customisation}\index{customisation!postnote}
%\paragraph{You want something \ldots\ but not `pp.'}
\paragraph{使用其他而非页码引导符} Some sources
don't use pages for pinpoint citations. For instance, suppose you are
citing a work which is normally referred to by section, not page. In
such a case, you can still use the convenience of having automatic
prefixes. What you need to do is to set the pagination field of the
source in the \texttt{.bib} file to `section'. A list of the types of
pagination recognised, and their results, is shown in table
\ref{pagination:types}.\index{pagination!in postnote}
\begin{margintable}
\begin{tabular}{lll}
  \toprule
  \textsf{pagination} & \textsf{singular} & \textsf{plural} \\
  \midrule
  \texttt{page}       & p.                & pp.             \\
  \texttt{column}     & col.              & cols.           \\
  \texttt{section}    & \S                & \S\S            \\
  \texttt{paragraph}  & par.              & pars.           \\
  \texttt{verse}      & v.                & vv.             \\
  \texttt{line}       & l.                & ll.             \\
  \texttt{none}                                             \\
  \bottomrule
\end{tabular}
\vspace{3pt}
\caption{Standard values for \texttt{pagination}\label{pagination:types}}
\end{margintable}


That’s fine so long as \biblatex\ recognises the pagination type. But if
it doesn't, you will get odd results. For instance, suppose we want
references in the form `Art. x', so that \cs{cite[2]\{key\}} gives us
`Art. 2'. If we enter the pagination as \texttt{article}, however, we
just get `p.', because \texttt{article} is not a
pagination type that \biblatex\ knows.

How can we `teach' \biblatex\ a new pagination type? In this example,
we simply need to define two bibliography strings---article and
articles:

\begin{verbatim}
\NewBibliographyString{article, articles}
\DefineBibliographyStrings{american}{% or your language
   article  = {Art\adddot}
   articles = {Arts\adddot}}
\end{verbatim}
Now all will work as intended.

%\paragraph{Modifying standard prefixes}
\paragraph{修改标准引导符} To change the standard
prefixes, you can use essentially the same trick -- only you don't
need to define the relevant bibliography strings because they already
exist. So, for instance, if you wanted the abbreviation for page to be
“pg”, you could redefine the relevant strings as follows

\begin{verbatim}
\DefineBibliographyStrings{american}{% or your language
  page  = {pg},
  pages = {pgs}}
\end{verbatim}
\indexstop{citation!postnote}

%\section{Multiple citations}
\section{多重引用}

\indexstart{citation!multiple}\csindex{cites}
You've probably worked out by now that even the standard citation
commands will handle multiple citations: \cs{cite\{key1, key2\}} will
produce `[1, 2]' or `Jones 1990; Smith 2010', or whatever is in
keeping with the style you have chosen.\marginpar{\footnotesize 多种引用的目的是为一个引用命令中的所有文献标注细节信息。
}

\begin{margintable}
\begin{tabular}{ll}
\toprule
\textsf{regular} & \textsf{multi-cite version} \\
\midrule
\cs{cite}        & \cs{cites} \\
\cs{footcite}    & \cs{footcites} \\
\cs{parencite}   & \cs{parencites} \\
\cs{autocite}    & \cs{autocites} \\
\cs{textcite}    & \cs{textcites} \\
\cs{Cite}        & \cs{Cites} \\
\cs{Parencite}   & \cs{Parencites} \\
\cs{Autocite}    & \cs{Autocites} \\
\cs{Textcite}    & \cs{Textcites}\\
\bottomrule
\end{tabular}
\vspace{3pt}
\caption{Multiple citation commands\label{multicites}}
\end{margintable}


This is fine; but it doesn't handle pre- and postnotes very nicely. If
we enter \cs{cite[see][11]\{key1,key2\}} we get something `[see 1;
2, p.~11]' --- which will only make sense if the page reference is
to the last work cited. For this purpose, \biblatex\ offers a
selection of commands tailored for multiple citations.

The pattern is set by the \cs{cites} command. If we type
\begin{center}
\verb|\cites[See][10]{key1}[cf][20]{key2}[30]{key3]|
\end{center}
we would get (in a numeric style)
\begin{center}
[see 1, p.\ 10; cf 2, p.\ 20; 3, p. 30]
\end{center}
or (in an alphabetic style)
\begin{center}
see Smith 1990, p.\ 10; cf Jones 2000, p.\ 20; Bloggs 2010, p.\ 30
\end{center}

It's not hard to see what's happening here: basically \biblatex\ is
interpreting \cs{cites} as a set of individual \cs{cite} commands,
each with its own pre- and/or postnote, but also understanding that
they form a unit (so, for instance, the numeric style places them in
one set of brackets).
\begin{center}
\ttfamily
\cs{cites}\colorbox{red!30}{[see][10]\{key1\}}\,%
\colorbox{green!30}{[cf][20]\{key2\}}\,%
\colorbox{blue!30}{[30]\{key3\}}
\end{center}

To make things even more exciting, you can also have a pre- or
postnote for the \emph{entire} multiple citation: you just place them
(before anything else) in parentheses:
\begin{center}
\ttfamily
\cs{cites}(see:)(among others)[10]{key 1}[20]{key 2}
\end{center}
giving
\begin{center}
see: Smith 1990, p.\ 10; Jones 2000, p.\ 20, among others
\end{center}

But there's one catch. As it scans forwards, the \cs{cites} command will
be `tricked' if it sees a square bracket (\texttt[) or a brace
(\texttt\{) which is not intended to start a citation. For instance,
in:
\begin{center}
\ttfamily
\cs{cites}\colorbox{red!30}{[10]\{key
  1\}}\,\colorbox{green!30}{[20]\{key2\}}
\colorbox{blue!30}{[`{}`the principal works'{}']}
\end{center}
a human can quickly see that the blue portion is not another
citation, but some text in brackets. But \biblatex\ won't appreciate
this (a space is not enough), and will complain that you have a
defective citation. The trick in such cases is to put \cs{relax} at
the end of the list:
\begin{center}
\verb|\cites[10]{key1}[20]{key2}\relax [`the principal works']|
\end{center}

All the main citation commands (and their capitalised forms) have
multiple versions, consistently named: see table \ref{multicites}.
\indexstop{citation!multiple}

%\section{Citation commands that \ldots\ don't}
\section{不标注的标注命令}

\csindex{nocite}
Finally we have to consider a small family of odd citation
commands. Normally the point of including a citation is to have
\emph{some sort} of information about the source in question. It
might, therefore, seem odd that there are citation commands which
don't print anything. But they are sometimes needed. For instance, you
might want to include a work in the bibliography without actually
printing any citation in the text. To do that you can use the \cs{nocite}
command. So \cs{nocite\{angled{key}\}} will place \emph{key} in the
bibliography without citing it. You can also use \cs{nocite\{*\}},
which will put \emph{every source} in your bibliography file into the
bibliography.

\csindex{mancite}
Another occasionally useful command is \cs{mancite}. You use this
command to tell \biblatex\ that you have manually cited a work. You
might like to do this if you want to weave the citation into the text
in a way that is too complex to achieve using the various citation
commands. It may be worth doing, because it will make sure that
\biblatex\ keeps track of internal housekeeping that matters. For
instance, suppose you had the following:
\begin{verbatim}
... is clear \parencite{jones}. Smith's `rebuttal' (see
above) hardly meets this point; but this is not Jones's
main objection \parencite[45]{jones}. ...
\end{verbatim}
If you were using a citation style which used `ibid', the second
reference to Jones would be printed as `ibid', since as far as
\biblatex\ is concerned there are two consecutive citations to it. But
the sentence describing Smith's rebuttal is a sort of implicit
citation, and you might think `ibid' ambiguous in this context. So you
could type
\begin{pseudoverb}
... is clear \cs{parencite}\{jones\}. Smith's `rebuttal' (see\\
above){\bfseries\cs{mancite}\{smith\}} hardly meets this point; but this is\\
not Jones's main objection \cs{parencite}[45]\{jones\}. ...
\end{pseudoverb}
Generally, however, it's better to avoid such implicit citations.

%%% Local Variables:
%%% coding: utf-8
%%% mode: LaTeX
%%% TeX-master: "biblatex-tutorial"
%%% End: