Syntactic Analyses for Parallel Grammars: Auxiliaries and Genitive NPs
Extracted source text ยท arXiv cmp-lg/9604013
% File modelcol96paper.tex
% 4 March 1996
% Contact: coling96@cst.ku.dk
% \documentstyle[colap]{article}
\documentstyle{article}
\def\nspl{\hbox{\hbox to 20pt{\hrulefill}}}
\def\buttline{\hbox{\hbox to 20pt{\ \hrulefill \ }}}
\newcommand{\mc}{\multicolumn}
\makeatletter
% File colap.sty
% 4 Mar 96
% Contact: coling96@cst.ku.dk
% This is the LaTeX style file for COLING-96. It is identical to the
% style file for EACL-95 and ANLP '94 except that the margins are
% adjusted for A4 paper. -- abney 13 Dec 94
% The ANLP '94 style file is a slightly modified
% version of the style used for AAAI and IJCAI, using some changes
% prepared by Fernando Pereira and others and some minor changes
% by Paul Jacobs.
% Papers prepared using the colsub.sty file and acl.bst bibtex style
% should be easily converted to final format using this style.
% (1) Submission information (\wordcount, \subject, and \makeidpage)
% should be removed.
% (2) \summary should be removed. The summary material should come
% after \maketitle and should be in the ``abstract'' environment
% (between \begin{abstract} and \end{abstract}).
% (3) Check all citations. This style should handle citations correctly
% and also allows multiple citations separated by semicolons.
% (4) Check figures and examples. Because the final format is double-
% column, some adjustments may have to be made to fit text in the column
% or to choose full-width (\figure*} figures.
% Place this in a file called colap.sty in the TeX search path.
% (Placing it in the same directory as the paper should also work.)
% Prepared by Peter F. Patel-Schneider, liberally using the ideas of
% other style hackers, including Barbara Beeton.
% This style is NOT guaranteed to work. It is provided in the hope
% that it will make the preparation of papers easier.
%
% There are undoubtably bugs in this style. If you make bug fixes,
% improvements, etc. please let me know. My e-mail address is:
% pfps@research.att.com
% Papers are to be prepared using the ``acl'' bibliography style,
% as follows:
% \documentstyle[colap]{article}
% \title{Title}
% \author{Author 1 \and Author 2 \\ Address line \\ Address line \And
% Author 3 \\ Address line \\ Address line}
% \begin{document}
% ...
% \bibliography{bibliography-file}
% \bibliographystyle{acl}
% \end{document}
% Author information can be set in various styles:
% For several authors from the same institution:
% \author{Author 1 \and ... \and Author n \\
% Address line \\ ... \\ Address line}
% if the names do not fit well on one line use
% Author 1 \\ {\bf Author 2} \\ ... \\ {\bf Author n} \\
% For authors from different institutions:
% \author{Author 1 \\ Address line \\ ... \\ Address line
% \And ... \And
% Author n \\ Address line \\ ... \\ Address line}
% To start a seperate ``row'' of authors use \AND, as in
% \author{Author 1 \\ Address line \\ ... \\ Address line
% \AND
% Author 2 \\ Address line \\ ... \\ Address line \And
% Author 3 \\ Address line \\ ... \\ Address line}
% If the title and author information does not fit in the area allocated,
% place \setlength\titlebox{<new height>} after
% \documentstyle[aclap]{article},
% where <new height> can be something larger than 2.25in
\typeout{Conference Style for COLING-96 -- released 4 March 1996}
% NOTE: Some laser printers have a serious problem printing TeX output.
% These printing devices, commonly known as ``write-white'' laser
% printers, tend to make characters too light. To get around this
% problem, a darker set of fonts must be created for these devices.
%
% Physical page layout - slightly modified from IJCAI by pj
% \setlength\topmargin{0.0in} \setlength\oddsidemargin{-0.0in}
% \setlength\textheight{9.0in} \setlength\textwidth{6.5in}
% \setlength\columnsep{0.25in} \newlength\titlebox \setlength\titlebox{2.25in}
% \setlength\headheight{0pt} \setlength\headsep{0pt}
% \setlength\footheight{0pt} \setlength\footskip{0pt}
% \thispagestyle{empty} \pagestyle{empty}
% \flushbottom \twocolumn \sloppy
% A4 version of page layout
\setlength\topmargin{-0.6cm} \setlength\oddsidemargin{-0cm}
\setlength\textheight{23.5cm} \setlength\textwidth{15.8cm}
\setlength\columnsep{0.25in} \newlength\titlebox \setlength\titlebox{2.25in}
\setlength\headheight{5pt} \setlength\headsep{0pt}
\setlength\footheight{0pt} \setlength\footskip{0pt}
\thispagestyle{empty} \pagestyle{empty}
\flushbottom \twocolumn \sloppy
% We're never going to need a table of contents, so just flush it to
% save space --- suggested by drstrip@sandia-2
\def\addcontentsline#1#2#3{}
% Title stuff, taken from deproc.
\def\maketitle{\par
\begingroup
\def\thefootnote{\fnsymbol{footnote}}
\def\@makefnmark{\hbox to 0pt{$^{\@thefnmark}$\hss}}
\twocolumn[\@maketitle] \@thanks
\endgroup
\setcounter{footnote}{0}
\let\maketitle\relax \let\@maketitle\relax
\gdef\@thanks{}\gdef\@author{}\gdef\@title{}\let\thanks\relax}
\def\@maketitle{\vbox to \titlebox{\hsize\textwidth
\linewidth\hsize \vskip 0.625in minus 0.125in \centering
{\Large\bf \@title \par} \vskip 0.2in plus 1fil minus 0.1in
{\def\and{\unskip\enspace{\rm and}\enspace}%
\def\And{\end{tabular}\hss \egroup \hskip 1in plus 2fil
\hbox to 0pt\bgroup\hss \begin{tabular}[t]{c}\bf}%
\def\AND{\end{tabular}\hss\egroup \hfil\hfil\egroup
\vskip 0.25in plus 1fil minus 0.125in
\hbox to \linewidth\bgroup\large \hfil\hfil
\hbox to 0pt\bgroup\hss \begin{tabular}[t]{c}\bf}
\hbox to \linewidth\bgroup\large \hfil\hfil
\hbox to 0pt\bgroup\hss \begin{tabular}[t]{c}\bf\@author
\end{tabular}\hss\egroup
\hfil\hfil\egroup}
\vskip 0.3in plus 2fil minus 0.1in
}}
\renewenvironment{abstract}{\centerline{\large\bf
Abstract}\vspace{0.5ex}\begin{quote}}{\par\end{quote}\vskip 1ex}
% bibliography
\def\thebibliography#1{\section*{References}
\global\def\@listi{\leftmargin\leftmargini
\labelwidth\leftmargini \advance\labelwidth-\labelsep
\topsep 1pt plus 2pt minus 1pt
\parsep 0.25ex plus 1pt \itemsep 0.25ex plus 1pt}
\list {[\arabic{enumi}]}{\settowidth\labelwidth{[#1]}\leftmargin\labelwidth
\advance\leftmargin\labelsep\usecounter{enumi}}
\def\newblock{\hskip .11em plus .33em minus -.07em}
\sloppy
\sfcode`\.=1000\relax}
\def\@up#1{\raise.2ex\hbox{#1}}
% most of cite format is from aclsub.sty by SMS
% don't box citations, separate with ; and a space
% also, make the penalty between citations negative: a good place to break
% changed comma back to semicolon pj 2/1/90
% \def\@citex[#1]#2{\if@filesw\immediate\write\@auxout{\string\citation{#2}}\fi
% \def\@citea{}\@cite{\@for\@citeb:=#2\do
% {\@citea\def\@citea{;\penalty\@citeseppen\ }\@ifundefined
% {b@\@citeb}{{\bf ?}\@warning
% {Citation `\@citeb' on page \thepage \space undefined}}%
% {\csname b@\@citeb\endcsname}}}{#1}}
% don't box citations, separate with ; and a space
% Replaced for multiple citations (pj)
% don't box citations and also add space, semicolon between multiple citations
\def\@citex[#1]#2{\if@filesw\immediate\write\@auxout{\string\citation{#2}}\fi
\def\@citea{}\@cite{\@for\@citeb:=#2\do
{\@citea\def\@citea{; }\@ifundefined
{b@\@citeb}{{\bf ?}\@warning
{Citation `\@citeb' on page \thepage \space undefined}}%
{\csname b@\@citeb\endcsname}}}{#1}}
% Allow short (name-less) citations, when used in
% conjunction with a bibliography style that creates labels like
% \citename{<names>, }<year>
%
\let\@internalcite\cite
\def\cite{\def\citename##1{##1, }\@internalcite}
\def\shortcite{\def\citename##1{}\@internalcite}
\def\newcite{\leavevmode\def\citename##1{{##1} (}\@internalciteb}
% Macros for \newcite, which leaves name in running text, and is
% otherwise like \shortcite.
\def\@citexb[#1]#2{\if@filesw\immediate\write\@auxout{\string\citation{#2}}\fi
\def\@citea{}\@newcite{\@for\@citeb:=#2\do
{\@citea\def\@citea{;\penalty\@m\ }\@ifundefined
{b@\@citeb}{{\bf ?}\@warning
{Citation `\@citeb' on page \thepage \space undefined}}%
\hbox{\csname b@\@citeb\endcsname}}}{#1}}
\def\@internalciteb{\@ifnextchar [{\@tempswatrue\@citexb}{\@tempswafalse\@citexb[]}}
\def\@newcite#1#2{{#1\if@tempswa, #2\fi)}}
\def\@biblabel#1{\def\citename##1{##1}[#1]\hfill}
%%% More changes made by SMS (originals in latex.tex)
% Use parentheses instead of square brackets in the text.
\def\@cite#1#2{({#1\if@tempswa , #2\fi})}
% Don't put a label in the bibliography at all. Just use the unlabeled format
% instead.
\def\thebibliography#1{\vskip\parskip%
\vskip\baselineskip%
\def\baselinestretch{1}%
\ifx\@currsize\normalsize\@normalsize\else\@currsize\fi%
\vskip-\parskip%
\vskip-\baselineskip%
\section*{References\@mkboth
{References}{References}}\list
{}{\setlength{\labelwidth}{0pt}\setlength{\leftmargin}{\parindent}
\setlength{\itemindent}{-\parindent}}
\def\newblock{\hskip .11em plus .33em minus -.07em}
\sloppy\clubpenalty4000\widowpenalty4000
\sfcode`\.=1000\relax}
\let\endthebibliography=\endlist
% Allow for a bibliography of sources of attested examples
\def\thesourcebibliography#1{\vskip\parskip%
\vskip\baselineskip%
\def\baselinestretch{1}%
\ifx\@currsize\normalsize\@normalsize\else\@currsize\fi%
\vskip-\parskip%
\vskip-\baselineskip%
\section*{Sources of Attested Examples\@mkboth
{Sources of Attested Examples}{Sources of Attested Examples}}\list
{}{\setlength{\labelwidth}{0pt}\setlength{\leftmargin}{\parindent}
\setlength{\itemindent}{-\parindent}}
\def\newblock{\hskip .11em plus .33em minus -.07em}
\sloppy\clubpenalty4000\widowpenalty4000
\sfcode`\.=1000\relax}
\let\endthesourcebibliography=\endlist
\def\@lbibitem[#1]#2{\item[]\if@filesw
{ \def\protect##1{\string ##1\space}\immediate
\write\@auxout{\string\bibcite{#2}{#1}}\fi\ignorespaces}}
\def\@bibitem#1{\item\if@filesw \immediate\write\@auxout
{\string\bibcite{#1}{\the\c@enumi}}\fi\ignorespaces}
% sections with less space
\def\section{\@startsection {section}{1}{\z@}{-2.0ex plus
-0.5ex minus -.2ex}{1.5ex plus 0.3ex minus .2ex}{\large\bf\raggedright}}
\def\subsection{\@startsection{subsection}{2}{\z@}{-1.8ex plus
-0.5ex minus -.2ex}{0.8ex plus .2ex}{\normalsize\bf\raggedright}}
\def\subsubsection{\@startsection{subsubsection}{3}{\z@}{1.5ex plus
0.5ex minus .2ex}{0.5ex plus .2ex}{\normalsize\bf\raggedright}}
\def\paragraph{\@startsection{paragraph}{4}{\z@}{1.5ex plus
0.5ex minus .2ex}{-1em}{\normalsize\bf}}
\def\subparagraph{\@startsection{subparagraph}{5}{\parindent}{1.5ex plus
0.5ex minus .2ex}{-1em}{\normalsize\bf}}
% Footnotes
\footnotesep 6.65pt %
\skip\footins 9pt plus 4pt minus 2pt
\def\footnoterule{\kern-3pt \hrule width 5pc \kern 2.6pt }
\setcounter{footnote}{0}
% Lists and paragraphs
\parindent 1em
\topsep 4pt plus 1pt minus 2pt
\partopsep 1pt plus 0.5pt minus 0.5pt
\itemsep 2pt plus 1pt minus 0.5pt
\parsep 2pt plus 1pt minus 0.5pt
\leftmargin 2em \leftmargini\leftmargin \leftmarginii 2em
\leftmarginiii 1.5em \leftmarginiv 1.0em \leftmarginv .5em \leftmarginvi .5em
\labelwidth\leftmargini\advance\labelwidth-\labelsep \labelsep 5pt
\def\@listi{\leftmargin\leftmargini}
\def\@listii{\leftmargin\leftmarginii
\labelwidth\leftmarginii\advance\labelwidth-\labelsep
\topsep 2pt plus 1pt minus 0.5pt
\parsep 1pt plus 0.5pt minus 0.5pt
\itemsep \parsep}
\def\@listiii{\leftmargin\leftmarginiii
\labelwidth\leftmarginiii\advance\labelwidth-\labelsep
\topsep 1pt plus 0.5pt minus 0.5pt
\parsep \z@ \partopsep 0.5pt plus 0pt minus 0.5pt
\itemsep \topsep}
\def\@listiv{\leftmargin\leftmarginiv
\labelwidth\leftmarginiv\advance\labelwidth-\labelsep}
\def\@listv{\leftmargin\leftmarginv
\labelwidth\leftmarginv\advance\labelwidth-\labelsep}
\def\@listvi{\leftmargin\leftmarginvi
\labelwidth\leftmarginvi\advance\labelwidth-\labelsep}
\abovedisplayskip 7pt plus2pt minus5pt%
\belowdisplayskip \abovedisplayskip
\abovedisplayshortskip 0pt plus3pt%
\belowdisplayshortskip 4pt plus3pt minus3pt%
% Less leading in most fonts (due to the narrow columns)
% The choices were between 1-pt and 1.5-pt leading
\def\@normalsize{\@setsize\normalsize{11pt}\xpt\@xpt}
\def\small{\@setsize\small{10pt}\ixpt\@ixpt}
\def\footnotesize{\@setsize\footnotesize{10pt}\ixpt\@ixpt}
\def\scriptsize{\@setsize\scriptsize{8pt}\viipt\@viipt}
\def\tiny{\@setsize\tiny{7pt}\vipt\@vipt}
\def\large{\@setsize\large{14pt}\xiipt\@xiipt}
\def\Large{\@setsize\Large{16pt}\xivpt\@xivpt}
\def\LARGE{\@setsize\LARGE{20pt}\xviipt\@xviipt}
\def\huge{\@setsize\huge{23pt}\xxpt\@xxpt}
\def\Huge{\@setsize\Huge{28pt}\xxvpt\@xxvpt}
\makeatother
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%% tree-dvips.sty from csli.stanford.edu
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% how to run.
% \input tree.sty or \documentstyle[tree]{..}
% run through latex as you normally do
% remember to send it to a postscript printer
\makeatletter
\special{header=tree-dvips.pro}
\def\pscmd#1{\special{ps:@beginspec
/nodemargin \@int{\the\nodemargin}\space pt def %
\@int{\the\treelinewidth}\space pt setlinewidth %
\ifdim\dashlength=0pt [] 0 setdash%
/arrowwidth \@int{\the\arrowwidth}\space pt def %
/arrowlength \@int{\the\arrowlength}\space pt def %
\else [\@int{\the\dashlength}\space pt] 0 setdash \fi\space
#1 %
@endspec}}
% parameters in tree mode
\newdimen\nodemargin %
\nodemargin=2pt
\newdimen\treelinewidth
\treelinewidth=.3pt
\def\thinline{\treelinewidth=.3pt}
\newdimen\dashlength
% \makedash{dimen} makes dashes of dimen length if dimen = 0pt then
% solid line
\def\makedash#1{\dashlength=#1}
\newdimen\arrowwidth
\arrowwidth 4pt
\newdimen\arrowlength
\arrowlength 4pt
{\catcode`p=12 \catcode`t=12 \gdef\wnum#1pt{#1}}
\def\@int#1{\expandafter\wnum#1}
\def\node#1#2{\leavevmode
\setbox1=\hbox{#2}\pscmd{/#1 \@int{\the\wd1} \space pt
\@int{\the\ht1} \space pt
\@int{\the\dp1} \space pt
node}\box1\relax}
% A node of 0pt width and height
% \nodepoint{nodename}[horizontal displacement][vertical displacement]
% default is current point
\def\nodepoint#1{\@ifnextchar [{\@nodepoint{#1}}{\@nodepoint{#1}[0pt][0pt]}}
\def\@nodepoint#1[#2][#3]{{\@tempdima=#2 \@tempdimb=#3%
\pscmd{/nodemargin 0 def %
/#1 \@int{\the\@tempdima}\space pt \@int{\the\@tempdimb}\space pt %
rmoveto 0 0 0 node}}}
\newtoks\pos@t \pos@t={top}
\newtoks\pos@b \pos@b={bottom}
\newtoks\pos@l \pos@l={left}
\newtoks\pos@r \pos@r={right}
\newtoks\pos@tl \pos@tl={topleft}
\newtoks\pos@tr \pos@tr={topright}
\newtoks\pos@bl \pos@bl={bottomleft}
\newtoks\pos@br \pos@br={bottomright}
% \nodeconnect[fromloc]{fromnode}[toloc]{tonode}
\def\nodeconnect{\@ifnextchar [{\@nodeconnect}{\@nodeconnect[b]}}
\def\@nodeconnect[#1]#2{\@ifnextchar [{\@@nodeconnect[#1]{#2}}{\@@nodeconnect[#1]{#2}[t]}}
\def\@@nodeconnect[#1]#2[#3]#4{\pscmd{%
/#2 getnode node\the\csname pos@#1\endcsname \space%
/#4 getnode node\the\csname pos@#3\endcsname\space nodeconnect}}
% \nodecurve[fromloc]{fromnode}[toloc]{tonode}{depthfrom}[depthto]
\def\nodecurve{\@ifnextchar [{\@nodecurve}{\@nodecurve[b]}}
\def\@nodecurve[#1]#2{\@ifnextchar [{\@@nodecurve[#1]{#2}}{\@@nodecurve[#1]{#2}[t]}}
\def\@@nodecurve[#1]#2[#3]#4#5{\@ifnextchar [{\@@@nodecurve[#1]{#2}[#3]{#4}{#5}}%
{\@@@nodecurve[#1]{#2}[#3]{#4}{#5}[#5]}}
\def\@@@nodecurve[#1]#2[#3]#4#5[#6]{\@tempdima=#5%
\@tempdimb=#6%
\pscmd{/depth \@int{\the\@tempdimb} \space pt def %
/#4 \the\csname pos@#3\endcsname cur\space %
/depth \@int{\the\@tempdima} \space pt def %
/#2 \the\csname pos@#1\endcsname cur\space %
nodecurve}}
% above is well tested below is not
\newif\iftransparent
\def\nodetriangle#1#2{\pscmd{/#1 /#2 nodetriangle}}
\makeatother
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% end tree-dvips.sty
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Lingmacros
% include
% \enumsentence, \eenumsentence, \ex
% \shortex, \shortexnt (\shortexdt no longer needed)
% this file created 1/23/89 by Emma Pease, CSLI
% modified 4/1/91 by Emma Pease
\makeatletter
\newcounter{enums}
% \widelabel is defined for use when the width of the enumsentence or
% eenumsentence number is wider than 20pt (default labelwidth). As a
% rule of thumb if your enumerated sentences are only 1 or 2 digit
% you don't need to worry. At 3 digits, set \widelabel to 2.78pt at 4
% digits set to 7.78pt.
\newdimen\widelabel
\widelabel=0pt
\def\enumsentence{\@ifnextchar[{\@enumsentence}% %]
{\refstepcounter{enums}\@enumsentence[(\theenums)]}}
\long\def\@enumsentence[#1]#2{\begin{list}{}{%
\advance\leftmargin by\widelabel \advance\labelwidth by \widelabel}
\item[#1] #2
\end{list}}
% The \ex definition
\newcounter{tempcnt}
\newcommand{\ex}[1]{\setcounter{tempcnt}{\value{enums}}%
\addtocounter{tempcnt}{#1}%
\arabic{tempcnt}}
% modified \@item command \unhbox\@tempboxa replaces \makelabel{#1}
% Used to get \eenumsentence to work correctly
\def\@item[#1]{\if@noparitem \@donoparitem
\else \if@inlabel \indent \par \fi
\ifhmode \unskip\unskip \par \fi
\if@newlist \if@nobreak \@nbitem \else
\addpenalty\@beginparpenalty
\addvspace\@topsep \addvspace{-\parskip}\fi
\else \addpenalty\@itempenalty \addvspace\itemsep
\fi
\global\@inlabeltrue
\fi
\everypar{\global\@minipagefalse\global\@newlistfalse
\if@inlabel\global\@inlabelfalse \hskip -\parindent \box\@labels
\penalty\z@ \fi
\everypar{}}\global\@nobreakfalse
\if@noitemarg \@noitemargfalse \if@nmbrlist \refstepcounter{\@listctr}\fi \fi
\setbox\@tempboxa\hbox{\makelabel{#1}}%
\global\setbox\@labels
\hbox{\unhbox\@labels \hskip \itemindent
\hskip -\labelwidth \hskip -\labelsep
\ifdim \wd\@tempboxa >\labelwidth
\box\@tempboxa
\else \hbox to\labelwidth {\unhbox\@tempboxa}\fi
\hskip \labelsep}\ignorespaces}
% for enumerated enumsentences. Internal enumeration is alpha if not
% otherwise defined.
% counter for items within \eenumsentence. (might use enumi instead?)
\newcounter{enumsi}
% \eenumsentence{\item[a.] A third \label{bar}example \toplabel{foo}
% \item[b.] A fourth \label{baz}example}
% Testing references \ref{foo}, \ref{bar}, and \ref{baz}.
% which produces
% (3) a. A third example
% b. A fourth example
% Testing references 3, 3a, and 3b.
% To ensure that labels in \eenumsentence come out right.
%
\def\theenumsi{\theenums\alph{enumsi}}
\newdimen\eeindent
\eeindent=15pt
% changes \leftmargin increased by \eeindent
% \labelwidth increased by \eeindent
% .. .. .leftmargin.. .. .. .. ..
% oldlabelwidth eeindent labelsep actual entry
% 20pt 15pt 5pt
% (enumlabel) eelabel
% eelabel
\def\@mklab#1{\hfil#1}
\def\enummklab#1{\hfil(\eelabel)\hbox to \eeindent{\hfil#1}}
\def\enummakelabel#1{\enummklab{#1}\global\let\makelabel=\@mklab}
\def\toplabel#1{{\edef\@currentlabel{\p@enums\theenums}\label{#1}}}
\def\eenumsentence{\@ifnextchar[{\@eenumsentence}% %]
{\refstepcounter{enums}\@eenumsentence[\theenums]}}
\long\def\@eenumsentence[#1]#2{\def\eelabel{#1}\let\holdlabel\makelabel%
\begin{list}{\alph{enumsi}.}{\usecounter{enumsi}%
\advance\leftmargin by \eeindent \advance\leftmargin by \widelabel%
\advance\labelwidth by \eeindent \advance\labelwidth by \widelabel%
\let\makelabel=\enummakelabel}
#2
\end{list}\let\makelabel\holdlabel}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% need to be modified?
% Format
% \shortex{4}{a & b & c & d}
% {z & y & x & w}
% {mnop}
% produces
% a b c d
% z y x w
% mnop
%
% \shortexdt takes two gloss lines
% \shortexnt takes no gloss lines
\def\shortex#1#2#3#4{\begin{tabular}[t]{@{}*{#1}{l@{\ }}}
#2\\ #3\\ \multicolumn{#1}{@{}l@{}}{\parbox{\linewidth}{#4}}
\end{tabular}}
%\def\shortexdt#1#2#3#4#5{\begin{tabular}[t]{@{}*{#1}{l@{\ }}}
%#2\\ #3\\ \multicolumn{#1}{@{}l@{}}{#4}\\
%\multicolumn{#1}{@{}l@{}}{#5}
%\end{tabular}}
\def\shortexnt#1#2#3{\begin{tabular}[t]{@{}*{#1}{l@{\ }}}
#2\\ #3
\end{tabular}}
\makeatother
%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% End of main lingmacros
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\newcommand{\bt}{\begin{tabular}}
\newcommand{\et}{\end{tabular}}
\newcommand{\st}{\shortex}
\newcommand{\ncon}{\nodeconnect}
\newcommand{\es}{\enumsentence}
\newcommand{\ees}{\eenumsentence}
\newcommand{\ua}{\uparrow}
\newcommand{\da}{\downarrow}
\newcommand{\ra}{\rightarrow}
% Note that we try to position the title about 1in from top of page.
% This allows slightly more room for title and author information
\title{\vspace{-0.5in}Syntactic Analyses for Parallel Grammars:
Auxiliaries and Genitive NPs}
\author{Miriam Butt \and Christian Fortmann \and Christian Rohrer \\
Institut f\"{u}r Maschinelle Sprachverarbeitung \\ Universit\"{a}t
Stuttgart \\
Azenbergstr. 12 \\ 70174 Stuttgart, Germany \\
\{mutt$|$fortmann$|$rohrer\}@ims.uni-stuttgart.de \\}
% Adjustments can be made if title and author information does not fit in
% the space provided - see comments on \titlebox in style file
\begin{document}
\maketitle
\vspace{-0.5in}
\begin{abstract}
This paper focuses on two disparate aspects of German syntax from
the perspective of parallel grammar development. As part of a
cooperative project, we present an innovative approach to
auxiliaries and multiple genitive NPs in German. The LFG-based
implementation presented here avoids unnessary structural complexity
in the representation of auxiliaries by challenging the traditional
analysis of auxiliaries as raising verbs. The approach developed
for multiple genitive NPs provides a more abstract, language
independent representation of genitives associated with nominalized
verbs. Taken together, the two approaches represent a step towards
providing uniformly applicable treatments for differing languages,
thus lightening the burden for machine translation.
\end{abstract}
\section{Introduction}
Within the cooperative parallel grammar project {\sc pargram}
(IMS-Stuttgart, Xerox-Palo Alto, Xerox-Grenoble), the analysis and
representation of structures in the grammars must be viewed from a
more global perspective than that of the individual languages (German,
English, French). One major goal of {\sc pargram} is the development
of broad coverage grammars which are also modular and easy to
maintain. Another major goal is the construction of {\em parallel}
analyses for sentences of the same type in German, English, and
French. If this can be achieved, the problem faced by machine
translation (MT) could be greatly reduced. Due to the recent
development of a faster and more powerful version of the LFG
(Lexical-Functional-Grammar) based {\em Grammar Writer's Workbench}
(Kaplan and Maxwell 1993) at Xerox, the implementation of a
linguistically adequate, broad coverage grammar appears viable. Given
the flexible projection-based architecture of LFG (Dalrymple et
al.~1995) and the MT approach presented in Kaplan et
al.~(1989),\footnote{See also Sadler et al.~(1990), Sadler and
Thompson (1991), Kaplan and Wedekind (1993), Butt (1994) for further
work on MT within LFG.} a robust MT system is already in place.
In this paper, we concentrate on two issues within the broader
perspective of {\sc pargram}: the treatment of auxiliaries and the
transparent representation of multiple genitive NPs in German. These
phenomena represent two areas for which generally accepted proposals
exist, but whose implementation in the context of parallel grammar
development throws up questions as to their wider, crosslinguistic,
feasibility. With respect to auxiliaries, the standard {\em raising}
approach that is usually adopted yields undesirable structural
complexity and results in idiosyncratic, language particular analyses
of the role of auxiliaries. With regard to genitive NPs, the standard
analysis for German yields structures which are too ambiguous for a
succesful application of machine translation. The following sections
present a solution in that morphological wellformedness
conditions are stated at a separate component, the {\em morphology
projection}. Furthermore, a representation of argument structure is
implemented that is related to, but not identical to the
representation of grammatical functions. Language particular
idiosyncratic requirements are thus separated out from the language
universal information required for further semantic interpretation, or
machine translation.
\section{The Formalism}
The architecture of LFG assumed here is the ``traditional''
architecture described in Bresnan (1982), as well as the newer
advances within LFG (Dalrymple et al.~, 1995). A grammar is viewed as
a set of {\em correspondences} expressed in terms of {\em projections}
from one level of representation to another. Two fundamental levels
of representations within LFG are the c(onstitutent)-structure and the
f(unctional)-structure. The c-structure encodes idiosyncratic phrase
structural properties of a given language, while the f-structure
provides a language universal representation of grammatical functions
(e.g., {\sc subj}ect, {\sc obj}ect), complementation, tense, binding,
etc. The correspondence between c-structure and f-structure is
not onto or one-to-one, but many-to-one, allowing an abstraction over
idiosyncratic c-structure properties of a language (e.g.,
discontinuous constituents).
In addition, several proposals exploring possible representations of a
s(emantic)-structure have been made over the years (e.g. Halvorsen and
Kaplan (1988), Dalrymple et al.~ (1993)). As the
realization of a separate semantic component is only planned for the
latter stages within {\sc pargram}, no further discussion of possible
formalisms will take place here. It should be noted, however, that
rudimentary semantic information, such as argument structure
information (lexical semantics), is encoded within the f-structures in
order to facilitate transfer in some cases. A case in point is
presented in the section on German genitive NPs.
\section{Auxiliaries --- a flat approach}
\subsection{The Received Wisdom}
Auxiliaries have given rise to lively debates concerning their exact
syntactic status (e.g.~Chomsky (1957), Ross (1967), Pullum and Wilson
(1977), Akmajian et al.~(1979), Gazdar et al.~(1982)): are they simply
main verbs with special properties, or should they instantiate a
special category {\sc aux}? Within current lexical approaches
(Lexical-Functional-Grammar (LFG), Head-driven Phrase Structure
Grammar (HPSG)), auxiliaries (e.g.~{\em have}, {\em be}) and modals
(e.g.~{\em must}, {\em should}) are treated as {\em raising} verbs,
which are marked as special in some way: in HPSG through an [{\sc
aux}: +] feature (Pollard and Sag 1994), in LFG (Bresnan 1982) by a
difference in {\sc pred} value.\footnote{See Falk (1984) for an early
LFG treatment of `do' in line with that proposed here, and Abeill\'{e}
and Godard (1994) for a similar treatment in French.} However, newer
work within LFG (Bresnan 1995, T.H.~King 1995) has been moving away
from the raising approach towards an analysis where auxiliaries are
elements which contribute to the clause only tense/aspect, agreement,
or voice information, but not a subcategorization frame. This view is
also in line with approaches within GB (Government-Binding), which see
auxiliaries simply as possible instantiations of the functional
category I (see also Halle and Marantz (1993)).
The ``traditional'' treatment of auxiliaries in both HPSG (Pollard and Sag
1994) and LFG has its roots in Ross's (1967) proposal to treat
auxiliaries and modals on a par with main verbs.\footnote{The term
{\em auxiliary} has often been taken to subsume both modals and
elements such as {\em have} and {\em be}. However, the distinction
between the two is necessary not only semantically, but also
syntactically. In German and (some dialects of)
English modals can be stacked, while the distribution of auxiliaries
is more restricted. Also, assuming that semantic interpretation is
driven primarily off of the f-structure, the relative embedding of
modals must be preserved at that level in order to allow an
interpretation of their scope and semantic force.}\ In particular,
auxiliaries are treated as a subclass of raising verbs (e.g.~Pollard
and Sag (1994), Falk (1984)). For example, a simple sentence like (\ex{1})
would correspond to the c-structure and f-structure shown in (\ex{2})
and (\ex{3}), respectively. Note that the level of embedding in the
f-structure exactly mirrors the c-structure: each verbal element takes
a complement.
\es{\st{7}
{Der & Fahrer & wird & den & Hebel & gedreht & haben}
{the & driver & will & the & lever & turned & gave}
{`The driver will have turned the lever.'}}
\small
\vbox{
\es{\ }
\bt[t]{ccccccc}
& \node{z}{S} & \\[2ex]
\node{b}{NP} & & \node{a}{VP} \\ [2ex]
\node{c}{der Fahrer} & \node{d}{V[+aux]} & & \node{e}{VP} \\ [2ex]
& \node{g}{wird} & \node{f}{NP} & & \mc{3}{c}{\node{v}{VP}} \\ [2ex]
& & \node{h}{den Hebel} & & \mc{3}{c}{\node{i}{V'}} \\ [2ex]
& & & \node{j}{V'} & \node{k}{V[+aux]} \\ [2ex]
& & & \node{u}{V} \\ [2ex]
& & & \node{l}{gedreht} & \node{m}{haben} \\
\et
\ncon{z}{a} \ncon{z}{b}
\nodetriangle{b}{c} \ncon{a}{d} \ncon{a}{e}
\ncon{d}{g} \ncon{e}{f} \ncon{e}{v}
\nodetriangle{f}{h} \ncon{v}{i}
\ncon{i}{j} \ncon{i}{k}
\ncon{j}{u} \ncon{k}{m}
\ncon{u}{l}
}
\normalsize
\footnotesize
\vbox{
\es{\ }
\[ \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`wird}<\hbox{{\sc xc}}>\hbox{{\sc s}}\mbox{'} \\
\hbox{{\sc tense}} & \hbox{{\sc pres}} \\
\hbox{{\sc subj}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \node{a}{\hbox{`Fahrer'}} \\
\hbox{{\sc case}} & \hbox{{\sc nom}} \\
\hbox{{\sc gend}} & \hbox{{\sc masc}} \\
\hbox{{\sc num}} & \hbox{{\sc sg}} \\
\hbox{{\sc spec}} & \hbox{{\sc def}} \\
\end{array} \right] \\
\hbox{{\sc xcomp}} & \left[ \begin{array}{ll} \hbox{{\sc pred}} &
`\hbox{haben}<\hbox{{\sc xc}}>\hbox{{\sc s}}\mbox{'} \\
\hbox{{\sc subj}} & \node{b}{[\ \ \ ]} \\
\hbox{{\sc xcomp}} & \left[ \begin{array}{ll} \hbox{{\sc pred}} &
`\hbox{drehen}<\hbox{{\sc s}}, \hbox{{\sc o}}>\mbox{'} \\
\hbox{{\sc subj}} & \node{c}{[\ \ \ ]} \\
\hbox{{\sc obj}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`Hebel'} \\
\hbox{{\sc case}} & \hbox{{\sc acc}} \\
\hbox{{\sc gend}} & \hbox{{\sc masc}} \\
\hbox{{\sc num}} & \hbox{{\sc sg}} \\
\hbox{{\sc spec}} & \hbox{{\sc def}} \\
\end{array} \right]
\end{array} \right]
\end{array} \right]
\end{array} \right] \]
\nodecurve[r]{a}[r]{b}{3in}[1in]
\nodecurve[r]{b}[r]{c}{3in}[.7in]
}
\normalsize
The main reasons to treat auxiliaries as complement taking verbs in
English are: 1) an account of VP-ellipsis, VP-topicalization, etc.\
follows immediately; 2) restrictions on the nature of the verbal
complement (progressive, past participle, etc.) following the
auxiliary can be stated straightforwardly (Pullum and Wilson (1977),
Akmajian et al.~(1979), Gazdar et al.~(1982)). The latter point holds
for German as well, and in fact, without some sort of a hierarchical
structure, stating wellformedness conditions on a string of multiple
auxiliaries becomes wellnigh impossible in light of the greater
ordering possibilities granted by the flexible German word order.
There are also major reasons, however, for not adopting this analysis:
1) linguistic adequacy; 2) unmotivated structural complexity; 3)
non-parallel analyses for predicationally equivalent sentences.
Consider the French equivalent of (\ex{-2}) in (\ex{1}).
\es{\st{6}
{Le & conducteur & aura & tourn\'{e} & le & levier}
{the & driver & will have & turned & the & lever}
{`The driver will have turned the lever.'}}
As argued by Akmajian et al.~(1979), crosslinguistic evidence
indicates that elements bearing only tense, mood, or voice should
belong to a distinct syntactic category. In many languages, like
French or Japanese, the information carried by {\em will} (future), or
{\em have} (perfect) is realized morphologically rather than
periphrastically. The analysis in (\ex{0}) thus effectively claims
that there exists a deep difference in the predicational structure of
auxiliaries like {\em will} and {\em have} and the French {\em
aura}.\footnote{Note that {\em wird} `will' is often analyzed as a
modal in accordance with Vater (1975). However, the arguments
presented there are not conclusive.} This is not desirable from a
crosslinguistic point of view, nor is it helpful for MT.
\subsection{Alternative Implementation}
The approach adopted here is a {\em flat} analysis of auxiliaries at
f-structure ((\ex{1})).
\es{
\[ \left[ \begin{array}{ll}
\hbox{{\sc pred}} & `\hbox{drehen}<\hbox{{\sc subj}}, \hbox{{\sc obj}} >\mbox{'} \\
\hbox{{\sc tense}} & \hbox{{\sc futperf}} \\
\hbox{{\sc subj}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \node{a}{\hbox{`Fahrer'}} \\
\hbox{{\sc case}} & \hbox{{\sc nom}} \\
\hbox{{\sc gend}} & \hbox{{\sc masc}} \\
\hbox{{\sc num}} & \hbox{{\sc sg}} \\
\hbox{{\sc spec}} & \hbox{{\sc def}} \\
\end{array} \right] \\
\hbox{{\sc obj}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`Hebel'} \\
\hbox{{\sc case}} & \hbox{{\sc acc}} \\
\hbox{{\sc gend}} & \hbox{{\sc masc}} \\
\hbox{{\sc num}} & \hbox{{\sc sg}} \\
\hbox{{\sc spec}} & \hbox{{\sc def}} \\
\end{array} \right] \\
\end{array} \right] \]
%\nodecurve[r]{a}[r]{b}{4in}[2in]
%\nodecurve[r]{b}[r]{c}{4in}[2in]
}
The auxiliaries {\em wird} `will' and {\em haben} `have' now only
contribute information as to the overall tense, but do not
subcategorize for complements. Structural phenomena like VP-ellipsis,
coordination, or topicalization can, however, still be accounted for
in terms of an appropriate embedding at c-structure (cf.~(\ex{-3})).
The role of auxiliaries in natural language is now adequately modeled,
in particular with respect to a more realistic treatment of tense
(compare (\ex{-2}) and (\ex{0})), as the French (\ex{-1}) has
essentially the same f-structure as (\ex{0}).\footnote{The
construction of the value for the composed tenses results from a
complex interaction between the lexical entries. Note that this
treatment does not as yet include a fine-grained represention of
tense and aspect. This is the subject of ongoing work. The
treatment presented here provides the basis needed for a thorough
crosslinguistic analysis of temporal and aspectual phenomena.}
However, the flat f-structure in (\ex{0}) provides no room for a
statement of selectional requirements, allowing massive overgeneration
(e.g.~nothing blocks the presence of two {\em haben} in (\ex{-4})).
Neither can the particular order of auxiliaries be regulated. Our
solution takes advantage of LFG's flexible projection-based
architecture by implementing a projection which models the
hierarchical selectional requirements of auxiliaries, yet does not
interfere with the subcategorizational properties of verbs, as would
be the case under a raising analysis.
\footnotesize
\vbox{
\es{\ }
\bt[t]{ccccc}
\multicolumn{3}{c}{\node{e}{VP}} \\ [2ex]
\node{f}{$\ua = \da$} &\multicolumn{3}{c}{\node{g}{$\ua = \da$}} \\
\node{h}{$\mu$ \sc{m}* $=\mu$*} &\multicolumn{3}{c}{\node{i}{($\mu$ \sc{m}* {\sc d}) $=\mu$*}} \\
\node{j}{AUX} &\multicolumn{3}{c}{\node{k}{VP}} \\ [2ex]
\node{jj}{wird} & &\multicolumn{2}{c}{\node{m}{$\ua = \da$}} \\
& \node{l}{($\ua$ {\sc xc}* {\sc gf}) $= \da$} &\multicolumn{2}{c}{\node{n}{$\mu$ \sc{m}* $=\mu$*}} \\
& \node{v}{NP} &\multicolumn{2}{c}{\node{o}{V$'$}} \\ [2ex]
& \node{vv}{den Hebel} \\ [2ex]
& & \node{p}{$\ua = \da$} & \node{q}{$\ua = \da$} \\
& & \node{r}{($\mu$ \sc{m}* {\sc d}) $=\mu$*} & \node{s}{$\mu$ \sc{m}* $=\mu$*} \\
& & \node{t}{V} & \node{u}{AUX} \\ [2ex]
& & \node{tt}{gedreht} & \node{uu}{haben} \\
\et
\ncon{e}{f} \ncon{e}{g}
\ncon{j}{jj} \ncon{k}{l} \ncon{k}{m}
\nodetriangle{v}{vv} \ncon{o}{p} \ncon{o}{q}
\ncon{t}{tt} \ncon{u}{uu}
}
\normalsize
In LFG, the flexible word order of German is handled via {\em
functional uncertainty}, which characterizes long-distance
dependencies without resorting to movement analyses (Netter (1988),
Zaenen and Kaplan (1995)). As in (\ex{0}), which illustrates our
alternative solution, functional uncertainty is represented by the
Kleene Star (*).\footnote{For space reasons, the {\sc xc} indicates
{\sc xcomp}, the {\sc d} a {\sc dep}.}\ The annotation on the NPs
indicates that they could fulfill the role of any possible grammatical
function (GF), e.g.~{\sc subj} or {\sc obj}, and that the level of
embedding ranges from zero to infinite. With every auxiliary
subcategorizing for an {\sc xcomp}, the two NPs could conceivably be
arguments of three different verbs: {\em wird}, {\em haben}, or {\em
gedreht}. Thus, the greater structural complexity unnecessarily
increases the search space for the determination of a verb's
arguments. In (\ex{0}), however, the m-structure is projected from the
c-structure parallel to the f-structure through annotations similar to
the usual f-structure annotations.\footnote{The annotation $\mu$ M* in
(\ex{0}) refers to the m-structure associated with the parent
c-structure node, and $\mu$* refers to the m-structure associated
with the daughter node. The more familiar $\uparrow$ and
$\downarrow$ of LFG are simply shorthand notations of the same idea,
but restricted to the projection from c-structure to f-structure:
$\uparrow = \phi$ \sc{m}*, $\downarrow = \phi$ *.}\ Statements about
``morphological'' dependents ({\sc dep}) are thus decoupled from
functional uncertainty: the relation of NP arguments to their
predicator now does not extend through various layers of artificial
structural complexity ({\sc xcomp}s). For VP-topicalization or
extraposition an unbounded long-distance dependency must still be
assumed. However, as the functional uncertainty path for auxiliaries
is distributed only over the m-structure of the verb complex (($\mu
\ua$ {\sc dep}*) $=\mu \da$), and does not involve the resolution of
the role of NP arguments, there are in fact differing paths of
functional uncertainty involved. The dependencies between predicators
and their arguments and auxiliaries and their dependents are thus
neatly factored out. The m-structure corresponding to the matrix VP in
(\ex{0}) is (\ex{1}). The desired flat f-structure resulting from the
usual $\ua$ and $\da$ annotations is as in (\ex{-1}).
\small
\es{\
\[ \left[ \begin{array}{ll}
\hbox{{\sc aux}} & + \\
\hbox{{\sc fin}} & + \\
\hbox{{\sc dep}} & \left[ \begin{array}{ll}
\hbox{{\sc aux}} & + \\
\hbox{{\sc fin}} & - \\
\hbox{{\sc vform}} & \hbox{{\sc base}} \\
\hbox{{\sc dep}} & \left[ \begin{array}{ll}
\hbox{{\sc fin}} & - \\
\hbox{{\sc vform}} & \hbox{{\sc perfp}} \\
\end{array} \right] \\
\end{array} \right] \\
\end{array} \right] \]
}
\normalsize
Like the f-structure, the m-structure is an attribute-value matrix.
It encodes language-specific information about idiosyncratic
constraints on morphological forms. The m-structure is not derived
from the f-structure. Rather, both representations are in
simultaneous correspondence with the c-structure. The following
(abbreviated) lexical entry exemplifies the pieces of information
needed. The disjunctive lexical entry for {\em wird} `will' in
(\ex{1}) takes the various combinatory possibilities of auxiliaries
and main verbs into account, and provides the appropriate tense
feature. For example, it requires that the embedded {\sc vform} be {\sc base},
and that there be no passive involved for a simple future like {\em
wird drehen}.
\es{\bt[t]{ll}
wird & AUX \\
% & ($\ua$ {\sc subj case}) $=$ {\sc nom} \\
% & ($\ua$ {\sc subj num}) $=$ {\sc sg} \\
& ($\mu$ \sc{m}* {\sc aux}) $= +$ \\
& \{ ($\mu$ \sc{m}* {\sc dep vform}) $=\!c$ {\sc base} \\
& \ \ ($\mu$ \sc{m}* {\sc dep dep vform}) $\neq$ {\sc perfp} \\
& \ \ ($\ua$ {\sc passive}) $\neq + $ \\
& \ \ ``simple future: wird drehen'' \\
& \ \ ($\ua$ {\sc tense}) = {\sc fut} \\
& $|$ \\
& \ \ ($\mu$ \sc{m}* {\sc dep vform}) $=\!c$ {\sc base} \\
& \ \ ($\mu$ \sc{m}* {\sc dep dep vform}) $=\!c$ {\sc perfp} \\
& \ \ ($\ua$ {\sc passive}) $\neq + $ \\
& \ \ ``future perfect: wird gedreht haben'' \\
& \ \ ($\ua$ {\sc tense}) $=$ {\sc futperf} \} \\
\et}
Features needed only to ensure language particular wellformedness are
no longer unified into the f-structure, cluttering a representation
that is meant to be language independent. In our analysis, only
features needed for further semantic interpretation, MT, or for the
expression of language universal syntactic generalizations are
represented at f-structure. For example, morphologically encoded
information like case, gender, or agreement is needed for statements
as to binding, predicate-argument relations, or the determination of
complex clause structures (given that agreement is generally
clause-bounded), and is therefore represented at f-structure.
Wellformedness conditions on adjective inflection or relative pronoun
agreement, however, can now be stated on the m-structure as
idiosyncratic, language particular information which can be ignored
for purposes of MT or semantic interpretation.
\section{Multiple Genitive NPs}
The differing surface realization of genitives within NPs in English
(preverbal NPs, postverbal PPs), French (postverbal PPs), and German
(preverbal NPs, postverbal PPs or NPs), poses a particular challenge
for a parallel grammar development project like {\sc pargram}. In
this paper, we suggest a treatment of multiple genitive NPs which not
only accounts for some restrictions on their distribution within
German, but also allows a language independent (universal)
representation, thus facilitating MT.
In general, the distribution of multiple NPs within NPs is an area of
German syntax which has not received a satisfactory account to date
(e.g., Pollard and Sag (1994), Bhatt (1990), Haider (1988)). In
German, nouns generally have at most one genitive which may occur in
a prenominal or postnominal position adjacent to the noun. Both kinds of
genitives have the same morphological shape. However, nominalizations
that are derived from a transitive verb allow for two genitives, one
in the prenominal, the other in the postnominal position.
The function of a genitive is generally expressed as indicating a
possessor: {\sc poss} within LFG. However, in the case of two
genitives, the assignment of two {\sc poss} values violates the
uniqueness-condition on f-structures and is furthermore insufficient
to distinguish the two differing kinds of genitives. We therefore
propose the utilization of two functions named {\sc gen1} and {\sc
gen2} in order to avoid association with any specific semantic role.
Furthermore, as genitives in the NP are generally optional, they are
taken to express no governed functions, i.e., they are not
subcategorized for by the noun. So {\sc gen1} and {\sc gen2} are
semantic functions in LFG on a par with, say, adjuncts. The
NP rule for German then is (\ex{1}).\footnote{Abstracting away from
bar-level considerations and further optional constituents, this
rule captures the restrictions that determine the dislocation of a
genitive in the matrix NP.}
\es{\bt[t]{lll}
NP $\ra$ & (\{DET:& $\ua = \da$ \\
& $|$ NP: & ($\ua$ {\sc gen1}) $ = \da$ \}) \\
& N: & $\ua = \da$ \\
& (NP: & ($\ua$ {\sc gen2}) $ = \da$) \\
\et
}
If the head-noun is not derived from, say, a verb, the single genitive
in either position is interpreted as a possessor. In case of a derived
nominal, however, a genitive is interpreted according to the thematic roles
assigned to the arguments of the verbal base. That means the functions
{\sc gen1} and {sc\ gen2} have to be linked to the appropriate roles.
Neither of the two functions is in principle restricted to any
specific role. But if both genitives are present they must be
interpreted according to a thematic role hierarchy.
As (\ex{1}) shows, if only one genitive is present, its prenominal
interpretation may be as agent or as patient. A postnominal (single)
genitive is interpreted as agent if the head noun is derived from an
intransitive, and as a patient/theme if derived from a transitive.
\ees{\item[a.]\st{2}
{Elisabeths & Lachen}
{Elisabeth-Gen & laughing}
{`Elisabeth's laughter'}
\item[b.]\st{2}
{Roms & Belagerung}
{Rome-Gen & siege}
{`Rome's siege'}}
However, if two genitives occur, as in (\ex{1}), the prenominal
genitive is restricted to an agent, and the postnominal one to patient.
This restriction must be encoded at some level, but does not follow
from the distiction between {\sc gen1} and {\sc gen2}, which are functions
that do not bear any semantic content on their own.
\es{\st{3}
{Karls & Behandlung & Peters}
{Karl-Gen & treatment & Peter-Gen}
{`Karl's treatment of Peter'}}
To obtain the correct linking, the argument structure of the verbal
base must be available. Since MT is based on f-structures within {\sc
pargram}, the argument structure has to be present at this level of
representation.\footnote{If a semantic or argument projection is
assumed (e.g., Halvorsen and Kaplan, 1988), this information may be
represented at another independent projection.}\ Nominalization is therefore
implemented as a morphologically driven process (lexical rule) which
eliminates {\sc subj} and {\sc obj} from the verb's subcategorization
frame and enters the verb's argument structure into the lexical entry
of the noun. This yields the optionality of genitives while
preserving the underlying semantics, as shown in (\ex{1}). The
association of {\sc gen1} and {\sc gen2} then is determined according
to a hierarchical order of arguments (Bresnan, 1995).
This approach also provides a means of handling certain cases of
categorial shift. For instance, in German temporal and conditional
adjuncts may be realized as PPs dominating an NP headed by a deverbal
noun. English does not have this option, but employs an adjunct-clause
instead. Here, the {\sc gen1} and {\sc gen2} functions of the German
f-structure have to be related correctly to the {\sc subj} and {\sc
obj} functions of the English f-structure.
\es{\bt{lll}
bei Karls & Darstellung & des Vorfalls \\
at Karl-Gen & report & the accident-Gen \\
mussten & alle & lachen \\
must-Past & all & laugh \\
\mc{3}{l}{`when Karl reported the accident} \\
\mc{3}{l}{everyone had to laugh'}
\et}
Here the linking of the {\sc gen1} and {\sc gen2} functions to the
appropriate thematic role in the German f-structure drives the
transfer of these functions to the {\sc subj} and {\sc obj} functions
of the English f-structure.
\es{
\[ \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`Darstellung'} \\ \\
\hbox{{\sc arg-str}} & \left[ \begin{array}{ll}
\hbox{{\sc arg1}} & \hbox{{\sc agent}} \\
\hbox{{\sc arg2}} & \hbox{{\sc theme}} \\
\end{array} \right] \\
\node{a}{\hbox{{\sc gen1}}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`Karl'} \\
\end{array} \right] \\
\node{b}{\hbox{{\sc gen2}}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`Vorfall'} \\
\end{array} \right] \\
\end{array} \right] \]
}
\es{
\[ \left[ \begin{array}{ll}
\hbox{{\sc pred}} & `\hbox{report}<\hbox{{\sc subj}}, \hbox{{\sc obj}}>\mbox{'} \\
\node{c}{\hbox{{\sc subj}}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`Karl'} \\
\end{array} \right] \\
\node{d}{\hbox{{\sc obj}}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`accident'} \\
\end{array} \right] \\
\end{array} \right] \]
}
\nodecurve[l]{a}[l]{c}{1in}[1in]
\nodecurve[l]{b}[l]{d}{1in}[1in]
Under this approach, languages now only differ with respct to the
categorial realisation of the function by case-marked NP or PP. Thus,
an application of this treatment not only provides an adequate
grammatical analysis of the NP in German, but also facilitates MT.
\section{Conclusion}
This paper has presented innovative approaches for two particular
syntactic phenomena: auxiliaries and multiple genitive NPs. The
analyses proposed allow the factorization of language particular,
idiosyncratic information. This results in a cleaner treatment of
auxiliaries by factoring out morphological wellformedness conditions,
and allows for the preservation of argument structure information in
cases like that of the German multiple genitive NP construction, where
syntactically dissimilar constructions express essentially the same
predicate-argument relations. As such, the work presented here can be
seen as a small but necessary step towards the realization of a broad
coverage grammar. In particular, the feasibility of developing
parallel grammars for differing languages is greatly increased through
the formulation of uniformly applicable, transparent analyses.
\section{Acknowledgments}
We would like to acknowledge Judith Berman, Mark Johnson, Ron Kaplan,
Mar\'{\i}a-Eugenia Ni\~{n}o and Annie Zaenen for the many valuable
discussions that served as input to this paper.
\begin{thebibliography}{}
\bibitem[\protect\citename{Akmajian \bgroup et al.\ \egroup }1979]{ak:aux}
Akmajian, A., S.~Steele, and T.~Wasow.
\newblock 1979.
\newblock The Category {AUX} in Universal Grammar.
\newblock {\em Linguistic Inquiry} 10(1):1--64.
\bibitem[\protect\citename{Bhatt }1990]{bhatt:np}
Bhatt, C. 1990.
\newblock {\em Die syntaktische Struktur der Nominalphrase im {D}eutschen}.
\newblock T\"{u}bingen: Narr.
\bibitem[\protect\citename{Berman and Frank }1996]{berman:lfg}
Berman, J., and A.~Frank. To appear.
\newblock {\em Deutsche und Franz\"{o}sische Syntax im Formalismus der LFG}.
\newblock T\"{u}bingen: Niemeyer.
\bibitem[\protect\citename{Bresnan }1995]{bresnan:str}
Bresnan, J. 1995.
\newblock {\em Lexical-Functional Syntax}.
\newblock Lecture Notes, ESSLI 7, Barcelona.
\bibitem[\protect\citename{Bresnan }1982]{bre:lfg}
Bresnan, J. (Ed.). 1982.
\newblock {\em The Mental Representation of Grammatical Relations}.
\newblock Cambridge, MA: MIT Press.
\bibitem[\protect\citename{Butt }1994]{butt:mt}
Butt, M. 1994.
\newblock Machine Translation and Complex Predicates
\newblock In {\em Proceedings of KONVENS 94}, 62---72.
\newblock Vienna, September.
\bibitem[\protect\citename{Dalrymple \bgroup et al.\ \egroup }1995]{dal:lfg-form}
Dalrymple, M., R.~Kaplan, J.~Maxwell, and A.~Zaenen. (Eds.). 1995.
\newblock {\em Formal Issues in Lexical-Functional Grammar}.
\newblock Stanford, CA: CSLI Publications.
\bibitem[\protect\citename{Dalrymple \bgroup et al.\ \egroup }1993]{dal:se}
Dalrymple, M., J.~Lamping, and V.~Saraswat. 1993.
\newblock {LFG} {S}emantics via {C}onstraints.
\newblock In {\em EACL 6}.
\bibitem[\protect\citename{Falk }1984]{falk:aux}
Falk, Y.
\newblock 1984.
\newblock The English Auxiliary System: a {L}exical-{F}unctional Analysis
\newblock {\em Language} 60(3):483--509.
\bibitem[\protect\citename{Haider }1988]{haider:np}
Haider, H.
\newblock 1988.
\newblock Die {S}truktur der deutschen {NP}.
\newblock {\em Zeitschrift f\"{u}r Sprachwissenschaft} 7(1):32--79.
\bibitem[\protect\citename{Halle and Marantz }1983]{hallemarantz:morph}
Halle, M., and A.~Marantz. 1993.
\newblock Distributed Morphology and the Pieces of Inflection.
\newblock In K.~Hale and S.J.~Keyser (Eds.), {\em The View from
Building 20}, 111--176.
\newblock Cambridge, Massachusetts: The MIT Press.
\bibitem[\protect\citename{Halvorsen and Kaplan }1988]{halvkap:sem}
Halvorsen, P.-K., and R.~Kaplan. 1988.
\newblock Projections and semantic description in
{L}exical-{F}unctional {G}rammar.
\newblock In {\em Proceedings of the International Conference on Fifth
Generation Computer Systems}, 1116--1122.
\bibitem[\protect\citename{Kaplan \bgroup et al.\ \egroup }1989]{kapetal:mt}
Kaplan, R., K.~Netter, J.~Wedekind, and A.~Zaenen. 1989.
\newblock Translation by structural correspondences.
\newblock In {\em EACL 4}, 272--281.
\newblock University of Manchester.
\bibitem[\protect\citename{Kaplan and Wedekind }1993]{kapwede:mt}
Kaplan, R., and J.~Wedekind. 1993.
\newblock Restriction and correspondence-based translation.
\newblock In {\em EACL 6}.
\newblock University of Utrecht, April.
\bibitem[\protect\citename{Kaplan and Maxwell }1993]{kapmax:gwb}
Kaplan, R., and J.~Maxwell. 1993.
\newblock Grammar Writer's Workbench
\newblock Ms., Xerox Corporation.
\bibitem[\protect\citename{King }1995]{thking:tofocus}
King, T.~H. 1995.
\newblock {\em Configuring Topic and Focus in {R}ussian}.
\newblock Stanford, California: CSLI Publications.
\bibitem[\protect\citename{Netter }1988]{netter:wo}
Netter, K. 1988.
\newblock Non-local dependencies and infinitival constructions in {G}erman.
\newblock In U.~Reyle, and C.~Rohrer(Eds.), {\em Natural language
parsing and linguistics theories}. Dordrecht: Reidel.
\bibitem[\protect\citename{Pollard and Sag}1994]{sag:hpsg}
Pollard, C., and I.~Sag. 1994.
\newblock {\em Head-Driven Phrase Structure Grammar}.
\newblock Chicago, IL: The University of Chicago Press.
\bibitem[\protect\citename{Pullum and Wilson }1977]{pullum:aux}
Pullum, G., and D.~Wilson.
\newblock 1977.
\newblock Autonomous Syntax and the Analysis of Auxiliaries.
\newblock {\em Language} 53(4):741--788.
\bibitem[\protect\citename{Ross }1967]{ross:syn}
Ross, J. 1967.
\newblock {\em Constraints on Variables in Syntax}.
\newblock PhD thesis, MIT.
\bibitem[\protect\citename{Sadler \bgroup et al.\ \egroup }1990]{sadetal:mt}
Sadler, L., I.~Crookston, D.~Arnold, and A.~Way. 1990.
\newblock {LFG} and translation.
\newblock In {\em Proceedings of the Third International Conference on
Theoretical and Methodological Issues in Machine Translation}, 11--13.
\newblock Linguistic Research Center, University of Texas at Austin.
\bibitem[\protect\citename{Sadler and Thompson }1991]{sadthompson:mt}
Sadler, L., and H.~Thompson. 1991.
\newblock Structural non-correspondence in translation.
\newblock In {\em EACL 5}, 293--298.
\newblock Berlin.
\bibitem[\protect\citename{Vater }1975]{vater:mod}
Vater, H. 1975.
\newblock {\em Werden} als {M}odalverb.
\newblock In Calbert, and H.~Vater (Eds.), {\em Aspekte der Modalit\"{a}t}.
T\"{u}bingen: Narr.
\end{thebibliography}
\end{document}
\es{
\[ \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`Behandlung'} \\
\hbox{{\sc arg-str}} & \left[ \begin{array}{ll}
\hbox{{\sc arg1}} & \hbox{{\sc agent}} \\
\hbox{{\sc arg2}} & \hbox{{\sc patient}} \\
\end{array} \right] \\
\hbox{{\sc gen1}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`Karl'} \\
\end{array} \right] \\
\hbox{{\sc gen2}} & \left[ \begin{array}{ll}
\hbox{{\sc pred}} & \hbox{`Peter'} \\
\end{array} \right] \\
\end{array} \right] \]
}