\documentclass{article}
\title{24.900: Fieldwork on the Mongolian language (Part 2)}
\author{Tony Kim}
\usepackage{tipa}
\usepackage{qtree}
\setlength{\parindent}{0pt}
\setlength\parskip{0.1in}
\setlength\topmargin{0in}
\setlength\headheight{0in}
\setlength\headsep{0in}
\setlength\textheight{8.5in}
\setlength\textwidth{6.5in}
\setlength\oddsidemargin{0in}
\setlength\evensidemargin{0in}

\pdfpagewidth 8.5in
\pdfpageheight 11in

\begin{document}
\maketitle

\section{Task 1: Basic clause structure}

I began the interview by asking Enky for the Mongolian equivalent of `John' and `Mary.' They are `Bat' (pronounced /\textipa{b2t}/) and `Dorma' (pronounced as it looks) respectively. We will be meeting these characters frequently in this assignment. In presenting this section, I was torn between the logic of presenting the sentences in order of complexity, but was overcome by the wish to present a coherent story.

Here we also give trees for some of the simpler sentences in the collection. It should be generally clear that, at least for unembedded sentences, Mongolian is a head-last language as Korean and Japanese.
\begin{enumerate}
	\item \textbf{`Bat likes scary books'}
	
	\begin{tabular}{llll}
		\textipa{b2t} & \textipa{emSkE} & \textipa{nom-d} & \textipa{de\*rte} \\
		Bat						& scary						& books*					& likes
	\end{tabular}
	
	* In the previous assignment we saw that the Mongolian word for `book' is /\textipa{nom}/. The suffix `-d' that follows in the above is not due to the plural form, but instead seems to be a marker for the object in the sentence. The actual marker evidently depends upon the verb of the sentence. When asked about a predictive rule, Enky was unable to formulate one.
	
	\item \textbf{`Bat likes scary books about ghosts.'}
	
	\begin{tabular}{llllll}
		\textipa{b2t} & \textipa{sunsni} & \textipa{tuxaj} & \textipa{emSkE} & \textipa{nom-d} & \textipa{de\*rte} \\
		Bat						& ghosts					 & about*					 & scary					 & books					 & likes
	\end{tabular}
	
	\Tree [.IP [.NP [.N' [.N \textipa{b2t} ] ] ] [.I' [.VP [.V' [.NP [.PP [.P' [.NP \textipa{sunsni} ] [.P \textipa{tuxaj} ] ] ] [.NP [.ADJ \textipa{emSkE} ] [.N' [.N \textipa{nom-d} ] ] ] ] [.V \textipa{de\*rte} ] ] ] [.I(-past) ] ] ]
	
	* The preposition `about' is treated differently in Mongolian from other prepositions such as `for', `into', `with', etc. In particular, `about' is special in that when formulating a prepositional phrase with the head `about', the lexical item for `about' does not attach to the object of the (c.f. Example 9 and 10, which are apparently more common forms for prepositional phrases). It seems to be a general rule that preposition morphemes are attached to the last word of the clause. `About' is an exception to this rule.
	
	\item \textbf{`Dorma knows Bat likes scary books.'}
	
	\begin{tabular}{llllll}
		\textipa{dorma} & \textipa{b2t-in} & \textipa{emSkE} & \textipa{nom-d} & \textipa{de\*rte} & \textipa{mit2n} \\
		Dorma					  & Bat							 & scary					 & books					 & likes						 & knows
	\end{tabular}
	
	\item \textbf{`Dorma hides under the bed.'}
	
	\begin{tabular}{lllll}
		\textipa{dorma} &	\textipa{orni}	& \textipa{d2\*r} & \textipa{noktitS} & \textipa{bEn} \\
		Dorma						& bed							& under*					& hides							& (I: present)
	\end{tabular}
	
	* Another exception to the preposition rule.
	
	\item \textbf{`Bat sees Dorma.'}
	
	\begin{tabular}{llll}
		\textipa{b2t} & \textipa{dorma-g} & \textipa{harS} & \textipa{bEn} \\
		Bat						& Dorma							& sees					 & (I: present)
	\end{tabular}
	
	\Tree [.IP [.NP \textipa{b2t} ] [.I' [.VP [.V' [.NP \textipa{dorma-g} ] [.V \textipa{harS} ] ] ] [.I(-past) \textipa{bEn} ] ] ]
	
	\item \textbf{`Bat saw Dorma.'}
	
	\begin{tabular}{lll}
		\textipa{b2t} & \textipa{dorma-g} & \textipa{harS-an} \\
		Bat						& Dorma							& saw*
	\end{tabular}
	
	* The suffix `-an' to the end of the verb is used to indicate past tense. I was informed that explicit tense morphemes (such as /\textipa{ben}/ `(I: present)') are not necessary, and that they can be replaced by these tense suffixes \emph{to verbs}. In the upcoming examples, we will see an example where an explicit future tense word is needed because the phrase does not terminate with a vowel.
	
	Furthermore, I was told that there are no past participles forms; e.g. there is no ``Bat has seen Dorma'' in Mongolian.
	
	\item \textbf{`Bat loves Dorma.'}
	
	\begin{tabular}{lll}
		\textipa{b2t} & \textipa{dorma-d} & \textipa{herte} \\
		Bat						& Dorma*							& loves
	\end{tabular}
	
	* In examples 5, 6, 7 we see clearly that the marker for the object of a sentence depends on the verb of the phrase (e.g. `-d' or `-g'). Enky was unable to give me a predictive rule.
	
	\item \textbf{`Bat wants to give Dorma a gift.'}
	
	\begin{tabular}{llllll}
		\textipa{b2t} & \textipa{dorma-d} & \textipa{bilIg} & \textipa{otSrIg} & \textipa{hustS} & \textipa{bEn} \\
		Bat						&	Dorma							& gift						& give						 & wants					 & (I: present)
	\end{tabular}
	
	\item \textbf{`Bat will marry Dorma if he can impress her.'}
	
	\begin{tabular}{llllll}
		\textipa{hirbe}	& \textipa{b2t}	& \textipa{dorma-d}	& \textipa{talagd-val}	& \textipa{tante} & \textipa{gilS-in} \\
		if							& Bat						& Dorma							& impress-if*						& her							& marry-(I: future)**
	\end{tabular}
	
	* In Mongolian, the `if' construction involves two morphemes which wrap the subordinate clause. I was unable to obtain any other construction that showed such splitting. I was similarly unable to obtain a sentence where the /\textipa{hirbe}/ and /\textipa{-val}/ occurs as a single unit. Hence, it is not clear if the process observed here is similar to that found in German verb splittings.
	
	** Note that, again, the tense of the sentence is marked with a suffix that attaches to the verb.
	
	\item \textbf{`Bat studied hard for a test.'}
	
	\begin{tabular}{llll}
		\textipa{b2t} & \textipa{shalgal-tant} & \textipa{sE} & \textipa{biltS-an} \\
		Bat						& test-for							 & hard					& studied-(I: past)
	\end{tabular}	
	
	\Tree [.IP [.NP \textipa{b2t} ] [.I' [.VP [.V' [.PP [.P' [.NP \textipa{shalgal} ] [.P \textipa{-tant} ] ] ] [.V' [.ADV \textipa{sE} ] [.V' [.V \textipa{biltS} ] ] ] ] ] [.I(+past) \textipa{-an} ] ] ]
	
	\item \textbf{`Bat married Dorma.'}
	
	\begin{tabular}{lll}
		\textipa{b2t} & \textipa{dorma-tai} & \textipa{gilS-an} \\
		Bat 					& Dorma-with					& married-(I: past)
	\end{tabular}
	
\end{enumerate}

\section{Task 2: Structure of complex NPs}

\begin{enumerate}
	\item \textbf{Enkhmunkh's father}*
	
	\begin{tabular}{ll}
		\textipa{enkmOnk-in} & av \\
		Enky-of 						 & father
	\end{tabular}
	
	\Tree [.NP [.N' [.PP [.P' [.NP \textipa{enkmOnk} ] [.P \textipa{-in} ] ] ] [.N' [.N \textipa{av} ] ] ] ]
	
	* During the interview, I actually requested ``father of Enky'', not ``Enky's father.'' I was informed, however, that both phrases would be mapped to the above Mongolian phrase. 
	
	\item \textbf{Girl with long har}
	
	\begin{tabular}{lll}
		\textipa{ort} & \textipa{us-t@}* & \textipa{ok2n} \\
		long					& hair-with				& girl
	\end{tabular}
	
	\Tree [.NP [.N' [.PP [.P' [.NP [.N' [.ADJ \textipa{ort} ] [.N' [.N \textipa{us} ] ] ] ] [.P \textipa{-t@} ] ] ] [.N' [.N \textipa{ok2n} ] ] ] ]
	
	* Refer to Sentence (11) of the previous section. During the interview, I observed that the morpheme for the preposition ``with'' assumes different forms in different sentences. Enky explained to me (or attempted, at the least!) that the forms of words are affected by the presence of ``masculine'' and ``feminine'' vowels. In particular, he claimed that the vowel `a' in (11) is an example of a masculine vowel, and hence the morpheme for ``with'' becomes /\textipa{-tai}/. On the other hand, he reported that `u' is a feminine vowel in this example, and therefore brings out the /\textipa{-t@}/ form.
	
	In this example and the previous, spoken Mongolian seems to suggest the possibility that the head of a prepositional phrase is ``attaching'' onto the noun object. However, I have drawn the structures here assuming that the underlying workings of Mongolian is identical to that of English. 
	
	\item \textbf{Girl with long hair who comes from America.}
	
	\begin{tabular}{lllll}
		\textipa{amerikas} & \textipa{irs-2n} & \textipa{ort} & \textipa{us-t@} & \textipa{ok2n} \\
		America						 & come-from				& long					& hair-with				& girl
	\end{tabular}
	
	\Tree [.NP [.N' [.PP [.P' [.NP \textipa{amerikas} ] [.P \textipa{irs-2n} ] ] ] [.N' [.PP [.P' [.NP [.N' [.ADJ \textipa{ort} ] [.N' [.N \textipa{us} ] ] ] ] [.P \textipa{-t@} ] ] ] [.N' [.N \textipa{ok2n} ] ] ] ] ]
	
	To simplify the structure, I may have misrepresented /\textipa{irs-2n}/ `come-from' by putting it as a single preposition. I do not feel equipped to tackle the potential complexity otherwise! As a matter of fact, I had originally asked for ``Girl with long hair from America'' rather than ``Girl with long hair \emph{who comes} from America.'' For some reason, Enky had trouble in translating the first form.
	
	I am also unsure which additions are modifiers or complements to the head.
	
	A bonus entry!
	
	\item \textbf{Bat loves the girl with long hair \emph{and} Dorma.}
	
	\begin{tabular}{lllllll}
		\textipa{b2t} & \textipa{ort} & \textipa{us-t@} & \textipa{ok2n} & \textipa{ba} & \textipa{dorma} & \textipa{herte} \\
		Bat						& long					& hair-with				& girl					 & and					& Dorma						& loves
	\end{tabular}
	
\end{enumerate}

\section{Task 3: Embedded declarative CPs}

One example of a sentence with an embedded clause (using the `if' complementizer) has been given as Sentence (10) of Section 1. We repeat it here and present another `if'-statement.

\begin{enumerate}
	\item \textbf{Bat will marry Dorma if he can impress her.}
	
	\begin{tabular}{llllll}
		\textipa{hirbe}	& \textipa{b2t}	& \textipa{dorma-d}	& \textipa{talagd-val}	& \textipa{tante} & \textipa{gilS-in} \\
		if							& Bat						& Dorma							& impress-if						& her							& marry-(I: future)
	\end{tabular}

	I find it difficult to draw a structure for this sentence because it seems like the main clause (``Bat will marry Dorma'') doesn't seem to have a subject!

	\item \textbf{Bat will be rich if he wins the lottery.}
	
	\begin{tabular}{lllllll}
		\textipa{hirbe} & \textipa{b2t} & \textipa{loDeri} & \textipa{hoj-vol} & \textipa{te\*r} & \textipa{bayan} & \textipa{bolno} \\
		if							&	Bat						& lottery					 & wins-if*					 & he							 & rich						 & will**
	\end{tabular}
		
	* A close examination with the previous example reveals that the morpheme for `if' is actually different in the two cases! (e.g. /\textipa{-val}/ vs. /\textipa{-vol}/.) This was explained by Enky as owing to vowel assimilation-like process.
	
	** It was noted by Enky that in this example, the explicit morpheme for the future tense is needed because the last word in the sentence otherwise (/\textipa{bayan}/ `rich') is an adjective and hence does not admit a suffix tense marker.
	
	I find it difficult to formulate the structure because it does not seem to follow the general head-last pattern. It is perhaps possible that the deep structure is as the following:
	
	\Tree [.CP [.C' (Placeholder) [.IP [.NP  \textipa{te\*r} ] [.I' [.VP [.V' [.CP \textipa{hirbe} [.C' [.IP [.NP \textipa{b2t} ] [.I' [.VP [.V' [.NP \textipa{loDeri} ] [.V \textipa{hoj} ] ] ] [.I(-pst) ] ] ] [.C \textipa{-vol} ] ] ] [.V' [.ADJ \textipa{bayan} ] [.V \textipa{bolno} ] ] ] ] [.I(-past) ] ] ] ] ]
	
	with a subsequent move operation on the CP clause to the outermost C' placeholder. However, I feel that it is difficult to postulate this explanation with only two examples.
	
	
\end{enumerate}

\section{Task 4: Yes/No questions}

\begin{enumerate}
	\item \textbf{Does Bat see Dorma?}
	
	\begin{tabular}{llll}
		\textipa{b2t} & \textipa{dorma-g} & \textipa{harS-an} & \textipa{ow}? \\
		Bat					  & Dorma							& see-past					& does?
	\end{tabular}
	
	\item \textbf{Does Bat like scary books?}
	
	\begin{tabular}{lllll}
		\textipa{b2t} & \textipa{emSke} & \textipa{nom-d} & \textipa{de\*rte} & \textipa{yu}? \\
		Bat						& scary						& books						& like							& does?
	\end{tabular}
	
	\item \textbf{Does the girl have long hair?}
	
	\begin{tabular}{lllll}
		\textipa{tir} & \textipa{ok2n} & \textipa{ort} & \textipa{us-t@} & \textipa{yu}? \\
		the						& girl					 & long					 & hair-with			 & does/is?
	\end{tabular}
	
	A more direct translation of this question may be ``Is the girl \emph{with} long hair?''
	
\end{enumerate}

Evidently, it seems to be a fairly simple business in Mongolian to construct ``yes/no'' questions. In the first two examples (for which I presented the declarative forms in Section 1) we see that we simply need to add a special complementizer at the end to construct the question form.

\section{Task 5: \emph{wh}-questions}

\begin{enumerate}
	\item \textbf{What did Bat study for?}
	
	\begin{tabular}{llll}
		\textipa{jund}	&	\textipa{b2t} & \textipa{biltSun} & \textipa{be}? \\
		What*						& Bat						& study     				& (a marker for a question) 
	\end{tabular}
	
	* In this sentence, it is permissible to exchange the order of the first two words. The sentence ``\textipa{b2t} \textipa{jund} \textipa{biltSun} \textipa{be}?'' is valid; and is in fact the first form that Enky produced. At this point, Enky remarked that in Mongolian one can ``get away with all sorts of rearrangements''! However, I have presented this \emph{wh}-first scenario to be more consistent with the other examples.
	
	\item \textbf{Who does Bat love?}
	
	\begin{tabular}{llll}
		\textipa{hind} & \textipa{b2t} & \textipa{he\*rte} & \textipa{ve}?* \\
		Who						 & Bat					 & love							 & (a marker for a question)
	\end{tabular}
		
	* Note the $b \rightarrow v$ swap. Enky was unable to explain the rule.
	
	\item \textbf{Why does Bat want to marry Dorma?}
	
	\begin{tabular}{lllllll}
		\textipa{jagat} & \textipa{b2t} & \textipa{dorma-te} & \textipa{gilSitig} & \textipa{huS} & \textipa{bEn} & \textipa{ve}? \\
		Why							& Bat						&	Dorma							 & marry							& want					& (I: present)  & (marker for question)
	\end{tabular}
	
\end{enumerate}

Similarly to the ``yes/no'' questions, the \emph{wh}-questions also require the use of a special complementizer which occurs at the end of the sentence. However, the \emph{wh} questions require an additional morpheme near the beginning of the sentence. While in these examples I have taken the \emph{wh}-words to be \emph{at} the beginning of the question for consistency, I was informed that the relative order of the first two words can be swapped.

\end{document}