Syntax of D: Argument-typing

Syntax is a set of rules governing the sentence structure in a given language. Often this is referred to with word ordering (sequencing) and structure of sentences. A lot of the world's languages follow a special word ordering where the subject comes first - whether SVO (subject-verb-object) or SOV (subject-object-verb).

Syntacticians are those who study syntax; they strive to discover the underlying syntactic rules that are common to all languages.

The DP Hypothesis

Debate: NP vs DP

Historically, the difference between NPs and DPs have been highly contested, but the Determiner Phrase (DP) hypothesis gained traction after Steven Abney postulated that the determiner (D) is actually the head of the noun phrase (NP). Until then, D was considered to occupy the specifier position of the noun it was projected from. There were several reasons against the use of DP, namely genitives, idiomatic expressions, absent determiners, morphological dependencies, left-branch phenomena; and syntactic and semantic parallelism. There is still an ongoing debate cross-linguistically about the use of NP vs. DP.

Defining DP

DP as a constituent

A DP is a type of phrase that has a head of a determiner, it usually comes before the noun when combined into a phrase. In phrases such as the cat, the is the determiner in the phrase and cat is the noun. When a determiner and a noun combine together, it can become a “determiner phrase”. In sentences and phrases where a noun does not explicitly show that it contains a determiner, DP hypothesis predicts the presence of a null determiner. For example, all argument Ns are DPs, consisting of [DP [D NP]], with D overt or covert.

Types of Ds

There are many different types of determiners. Some languages introduce their nominal structures using items from the following five groups: demonstratives, possessives, cardinal numerals, articles and quantifiers.

Type of determiner	Example
Demonstrative determiner	This, that, these, those
Possessive determiner	My, your, his, her, their, our
Cardinal numeral determiner	One, three, five hundred
Article determiner	A, an, the
Quantifier determiner	Many, some, few, any

Identifying DP types

As broad as the DP hypothesis can seem, what stood out when researching was the idea that it can be broken down into syntactic and semantic features and beyond. We specifically chose syntactic and semantic features which encompasses the three diagnostics: case, specificity and definiteness. Case lands under a syntactic correlate and Specificity and definiteness are for semantic features.

DP typing in English

Syntactic correlate of DP typing: case

Case for DPs is something that is dependent on the syntactic position in which the DP is located. However, in English specifically, case seems to only appear in certain pronouns and less so for determiners.

As mentioned, determining what case a DP identifies with, depends on where the DP occurs. Each DP is different, meaning that no two DPs appear in the exact same syntactic position in one sentence. The most common cases are nominative, accusative and genitive; their definitions are given below (these definitions are specific to a case that a DP gets - tying to the position in which the DP occurs in):

• Nominative often glossed as (NOM): subject of tensed T, so it's selected by tensed Ts.
• Accusative glossed as (ACC): complement of V, so it is a property selected by (certain) Vs.
• Genitive glossed as (GEN): subject of DP, so it is a property selected by the “possessive” D.

We know that in English, DP case is only visible to us on certain pronouns, but a handful of other languages in the world seem to show 'case’ on almost all of their DPs.

Standard Case Theory: According to Standard Case theory, there are two types of case -  structural and nonstructural case. However, it has been argued that nonstructural case further divides into 2 types - lexical case and inherent case. Case theory distinguishes two kinds of non-structural case, those being inherent case and lexical case.

Determining structural/non-structural case

Woolford highlights that the most common diagnostic test to determine whether something is under structural case or non-structural case is using the technique “Case Preservation under A-Movement (in voice raising as passive cannot be structural case). This technique states that if the case of an argument is kept the same (preserved) under the movement of A, this argument is said to have a non structural case. This means that those with structural cases will change its case after A-movement, and this change depends on what structural case is licensed in the position where this argument is moving towards.

2 kinds of nonstructural case:

1. Inherent case: case inherently associated with certain theta roles or theta positions
   1. Inherent case is also known as the “oblique case” where it usually depends on the ‘theta-positions’ as opposed to the structural cases. Inherent cases seem to happen more often and are more predictable.
2. Lexical case: lexically selected case, idiosyncratic
   1. Lexical Case is also known as the “idiosyncratic” and it is normally “lexically” selected by a lexical head.

A theta-role is the semantic relationship of an argument with its predicate, an argument that has been assigned a specific role i.e. agent and theme. In inherent cases, a theta-position is defined by where a theta-role is assigned in an A position.

Lexical and Inherent case as defined by Woolford is said to be in a “complementary distribution relationship with theta-positions; more specifically, the complementary distribution of lexical and inherent case are as follows:

• Lexical cases can occur on internal arguments as well as theme arguments, however they cannot occur on external/agentive arguments nor on the DP goal arguments (the ones that have been shifted).
• Inherent case can occur on external/agentive arguments and on the (shifted) DP goal arguments, however they cannot occur on internal arguments/theme arguments

To account for this complementary distribution between inherent case and lexical case, one needs to take a further look in depth to the reasoning behind why lexical and inherent case are in complementary distribution, as well as how non-structural case is licensed.

Woolford clarifies the following in her paper as she attempts to shed light on non-structural case licensing.

1. Lexical case is licensed only by lexical heads
2. Inherent case is licensed only by little/light v heads.

Semantic correlates of DP typing

Definiteness

Determiners (D) are used to show whether a noun/nominal is interpreted as definite/indefinite. It is a feature of semantics, and differs between languages. Definiteness is used for specificity, which will be further discussed below. These two ideas are usually taught early on in elementary school and can seem confusing at first, but most children are able to acquire the skills to correctly use it very quickly.

Indefinite determiners are used when the speaker wants to introduce an element to the listener that is only known to the speaker, and not the listener. Later on in the conversation, after the listener becomes familiar to the element, the speaker can change from using the indefinite determiner to the definite determiner.

Syntacticians have analyzed the definite article being the mechanism that binds NP variables.

A definite determiner, on the other hand, such as “the”, is used to select for a specific set, while indefinite determiners, such as “a” or “an”, are used in a more general sense, and do not specify a specific set. Definite determiners can be used in the singular context and plural contexts, whereas indefinite determiners are usually used in a singular context. Below is a table that contains a few examples and is helpful when learning about definite versus indefinite determiners.

Definite vs. Indefinite Determiners

	Definite	Indefinite
singular	The cookie	A cookie
plural	The cookies	Cookies

Sometimes determiners are shown openly (overt), but sometimes they can be silent (covert). For example, in the table above, the plural, indefinite phrase “cookies” has a silent determiner, while the singular, indefinite phrase as the overt determiner “a”.

Empty elements

Even with determiners, there can be places where we have empty/unpronounced elements when drawing out a syntactic tree. Not surprisingly though, definite and indefinite determiners are also able to take part in this. Empty definite determiners are said to only take NP complements headed by proper nouns but on the contrary, empty indefinite determiners only take NP complements headed by plural nouns.

Indefinite/definite determiners and complements

It’s likely that the definiteness of the determiners can impose restrictions on its NP complement, specifically those dealing with numbers. As mentioned above in the “empty elements” sub-section; (empty) definite determiners can only take NP complements headed by proper nouns and an (empty) indefinite determiner can only take NP complements headed by plural nouns.

Below is an example from the book Basic English Syntax:

Singular complement	Plural complement
A man	*a men
*both man	Both men

In addition, determiners as heads are able to project their properties to the sentence meaning that a plural indefinite determiner will be the head of a plural indefinite DP - and that a singular definite determiner will be the head of a singular definite DP.

Here are some examples to further clarify the idea between plural indefinite determiners being the head of plural indefinite DPs and singular definite determiners being heads of singular definite DPs.

There are some women in the store

*there is/are the woman/women in the store

The woman is in the store

The women are in the store

Specificity

While definiteness helps pick out one object from a group of possible objects, specificity relates elements that were previously established in the discourse. Namely, determiners are in place to indicate whether a noun is general or specific. Although they are often discussed in relation to each other, definiteness and specificity in themselves are considered to be distinct categories. This is something that is sensitive to speakers, no matter what language they speak, whether it is explicitly marked or not.

Specific Determiners

There are a handful of specific determiners, and each is used depending on the discourse context. We specify determiners when we believe/think that the listener (the person on the other side of the conversation) knows what we are talking about. Different determiners force different interpretations of the bare nouns they introduce.

Determiners that specify include the following:

Type of determiner	Example
Definite article	the
Demonstratives	this, that, these, those
Possessives	your, his/her, my, their

However, it is important to note that specificity is about more than just specific determiners. As mentioned previously, it is essential to make the distinction that specific doesn’t mean definite. This is because there is a structural difference between a projection that introduces specificity, and another projection where definiteness is introduced. There is a functional difference between the two. Syntacticians have commonly approached the analysis of specificity as a feature of indefinites, even though definite can be [± specific] as well. That is, it is possible for a definite DP to have a non-specific reading. Specific indefinites can be used in conjunction with proximate demonstratives (ex. this, these, those, etc.).

The following is a chart summarizing the possible readings taken from Ihsane and Puskas (2001):

	+[definite]	[-definite]
[+specific]	[+def, +spec]	[-def, +spec]
[non-specific]	[+def, non-spec]	[-def, non-spec]

Here are some more examples in English. Consider the following sentences:

(1a) I have eaten the cake

(1b) Chris has read the book.

In (1a), as shown above, the definite DP “the cake” can both get a specific interpretation, as well as a non-specific interpretation. The non-specific interpretation is made possible in the case that the referent of the DP “the cake” has not previously been established. This sentence would describe the event of the cake being eaten, with the cake being a non-specific cake. As we can see, despite the use of the definite determiner “the”, we still have a non-specific reading. This exemplifies the distinction between definiteness and specificity.

DP typing in Korean

Korean is an article-less language. Despite the fact that DP hypothesis is widely adapted cross-linguistically, it could not readily apply to Korean nominals as determiners are optional in Korean language. Many Korean linguists suggest that it is the NPs (bare or with numeral classifier) that play an important role in introducing arguments like case, definiteness and specificity. The studies that specifically deal with the entire structure of DP in Korean are limited. Therefore, the puzzles regarding how it is different for syntax of D in Korean to introduce arguments are left for future studies.

Below is an example of where a demonstrative determiner precedes a noun:

Ku	chaek
That	book

'That book.'

Basically, what is different from English is that in the DP structure of Korean, the demonstrative determiner ku can be optional but the noun head must not be optional. There is no restriction for any noun, chaek in this example, to mandatorily accompany a determiner, whereas in English a determiner should be present especially when the noun is singular and countable. Korean as an article-less and numeral classifier language, it is still considered grammatical even though a noun, regardless of whether it is count or mass and singular or plural, stands without a determiner.

Korean Determiners

Type	Korean	English Translation
Demonstrative	i	‘this’ (proximal)
	ku+	‘that’ (medial)
	jeo	‘that’ (distal)
Case	-ka, -i	Nominative
	-(l)ul	Accusative
	-uy	Genitive
	-ehgeh	Dative/‘to’
	-(n)un	Topical
Definite	ku+	‘the’
Indefinite	han/eo-nu	‘a’
	eo-ddeon/myeot	‘some’

ᐩ ku can both mean the medial demonstrative and the definite determiner.

Syntactic correlate of DP typing: case

In accordance with Abney’s DP approach to English NPs, Kim and Jeong suggest that Korean case markers belong to the head of DP, which then means that D is the locus of case features of Korean NPs. Given that Korean is head-final language, Jeong suggests the tree structure on the right:

Case markers in Korean occupy the D-heads.

However, when discussing the genitive and dative case markers in Korean language, things get more complex. Whether cases are DP-internally generated in Korean has been a popular debate in Korean language. Korean syntacticians argue that the role of D in languages like Korean and Japanese are vague. Jo proposed that Korean lacks a DP layer because the role of D (such as case marker) is divided and carried out by other elements in the sentence.

With an assumption that genitive mark fill the D-head position in English, 1(a) is ungrammatical because the extra determiner ‘the’ has no place to go. However, in Korean, the phrase 1(b) is grammatical even though both a genitive marker and a demonstrative determiner are used at the same time. This yields a hypothesis that the genitive case might not function as the head of D in all languages.

(1a) * Lucy's the book

(1b)

Lucy-uy	ku	chaek
Lucy-GEN	that	book

'That book of Lucy.'

Uy is the most typical genitive case that marks semantic arguments of relational nouns (e.g. eonni ‘older sister’, chingu ‘friend’, etc.) and deverbal nouns. Examples are shown below.

(2a)

Lucy-uy	eonni
Lucy-GEN	older sister

'Lucy's sister.'

(2b)

ilbon-uy	chimlyak
Japan-GEN	invasion

'The invasion of Japan.'

Aside from uy being a typical genitive case marker, it also attaches to other varieties of a prenominal constituents such as PPs and numerical classifiers. Examples in (3)a and (3)b show non-argument genitive phrases where the genitive phrase uy is not an argument of the head of nouns.

(3a)

hankuk-eulo-uy	yeohaeng
Korea-DIR-GEN	trip

‘a trip to Korea’

(3b)

se - myung - uy	namja
Three - Cl - GEN	man

‘Three men’

The examples in (2) are the argument genitives and (3) are the non-argument genitives. The former case takes the genitive marker uy as an option as it is still grammatical even without the genitive marker while the latter should take uy mandatorily as shown in (4). In other words, the uy in the argument genitive phrases like (2) are assumed to be the head of D in DP, while uy in (3) are assumed to be outside of domain of head nouns where uy-marked-constituents are modifiers rather than arguments.

(4a)

Lucy(-uy)	eonni
Lucy(-GEN)	older sister

‘Lucy’s sister’

(4b)

* se - myung (- uy)	namja
Three - Cl - GEN	man

‘Three men’

Knowing uy is optional in the argument genitive phrases, below example shows that the genitive marker on Lucy in (1)b Lucy-uy ku chaek was not a structural case marker as it cannot be omitted. This means uy following Lucy is not an argument genitive that is not the D-head but an adjunct modifying the DP.

(5)

* Lucy	ku	chaek
Lucy	that	book

* ‘That book of Lucy’ (intended meaning)

The tree on the right shows genitive uy as a PP adjunct to DP, and the tree structure is modified from Kang (2016):

uy as an adjunct PP to DP. It does not occupy a D-head in this case.

With these cases, it is challenging to treat all genitive phrases as a structural case marker that is assigned by the D-head.

In terms of the dative case, Lee (1997) proposed that the dative case -ehkeh is no different from the genitive case that -ehkek is not a case marker as well but rather a postposition which makes the dative construction a PP (at least when the verb is ‘give’). Below are the list of arguments that Lee (1997) suggested to account for the dative not being a case marker that is in the projection of D but a postposition that is in the projection of P:

(6)  Case drop: “A case marker can be dropped but a preposition cannot”.

(a)

John-i	Lucy-ehgeh	chaek-ul	ju-eot-da
John-NOM	Lucy-DAT	book-ACC	give-PST-DEC

‘John gave a book to Lucy.’

(b)

* John-i	Lucy	chaek-ul	ju-eot-da
John-NOM	Lucy	book-ACC	give-PST-DEC

* ‘John gave a book to Lucy.’ (intended meaning)

In this example, the dative cannot be removed as it generates an ill-formed sentence.

(7)  Case stacking: “Case stacking can apply to a postposition but not to a case marker”.

(a)

John-i	Lucy-ehgeh	chaek-ul	ju-eot-da
John-NOM	Lucy-DAT	book-ACC	give-PST-DEC

‘John gave a book to Lucy.’

(b)

John-i	Lucy-ehgeh-lul	chaek-ul	ju-eot-da
John-NOM	Lucy-DAT-ACC	book-ACC	give-PST-DEC

‘John gave a book to Lucy.’

Above examples show that the dative case -ehgeh shows other cases such as ACC can occur together.

(8)  Conjunction: “A postposition can mark the first of two conjuncts, but a case marker cannot”.

(a)

John-i	[Lucy-wa Sue]-ehgeh	chaek-ul	ju-eot-da
John-NOM	Lucy-and Sue-DAT	book-ACC	give-PST-DEC

‘John gave a book to Lucy and Sue.’

(b)

John-i	[Lucy-ehgeh-wa Sue]-ehgeh	chaek-ul	ju-eot-da
John-NOM	Lucy-DAT-and Sue-DAT	book-ACC	give-PST-DEC

‘John gave a book to Lucy and Sue.’

The examples demonstrate that eh-keh can mark the first NP Lucy.

To summarize, the genitive marker uy in Korean displays other circumstances that suggest its role as an adjunct PP to a DP rather than a structural case marker that is assigned by the D-head. And further evidence refuting the existence of true case markers in Korean comes from case drop, case stacking and conjunction as seen with dative case -ehgeh. The fact that case-suffixes -ka, -uy and -lul as well as postpositions -ehgeh and -uy occupy the same morphological position (i.e. they all attach to stem noun) makes the syntacticians cast doubt on treating them as structural case markers. The different distibutational circumstances for genitive & dative vs. nominative & accusative leave the syntacticians to further explore to what extent the D-heads take part in Korean language.

Semantic correlates of DP typing

Definiteness

In English, among other languages, DPs have functional elements such as case markers, determiners, quantifiers, and articles that denote definiteness and indefiniteness. Within a DP, a definite refers to an entity that is already familiar to both the speaker and the listener of a sentence. An indefinite refers to an entity that is newly introduced.

By contrast, Korean nouns do not require determiners. The most common form of an NP is a bare noun, where elements such as plurality and definiteness are inferred from the context of a sentence. NPs do not have indefinite or definite articles attached; instead, definiteness can be shown through classifiers, a required suffix for Korean nouns, or the demonstrative determiner ku.

There have been two agreed-upon qualities of definiteness in classifier languages such as Korean, where determiners are not required: the first is unique identifiability, and the second is familiarity. Unique identifiability means the entity referred to in a sentence can be identified, even if the participants are not previously familiar with the entity. Familiarity means the entity must refer to a concrete, or identifiable, entity.

Korean bare nouns adhere to the concept of unique identifiability under the following contexts, where unique identifiability typically appears in similar languages. In these contexts, definiteness can be shown through definite NPs. That is, sentences under these contexts are grammatical with the use of bare nouns, and definiteness can still be demonstrated. Bare nouns come with their restrictions as well, which are listed in the chart. Violations of these restrictions would require an overt determiner to show definiteness.

Unique identifiability in Korean bare nouns

Context	Explanation	Example
Contexts
Weather	NPs become bare nouns when describing phenomenons such as the sun or the sky.	Yeoseot-si-eh (*ku)hae-ka ji-n-da. Six-o’clock-LOC (the) sun-NOM set-PRES-DEC. ‘The sun sets at 6.’
Unique and common culture	NPs become bare nouns when describing common cultural or social entities that are uniquely known, such as popular figureheads.	Oh-nyeon-mada (*ku) daetongryeong-ul bbob-nun-da. Five-year-each (the) president-ACC elect-PRES-DEC. ‘We elect the president every four years.’
Immediate situations	NPs become bare nouns in immediate situations, where both the speaker and listener of a sentence share the same unique knowledge, such as when they are referring to a specific instance both of them are already familiar with.	Seuweteo eodi it-ni? Sweater where exist-INT? ‘Where is the/your sweater?’
Bridging	NPs become bare nouns if it has an antecedent, such as one that indicates possession. This is also known as associative anaphora.	Ce cha-nun penhophan-i eps-ese kyengchal-ey caphi-ess-ta. That car-TOP license plate-NOM not have-CONJ police-by caught-PAST-DEC. ‘That car was caught by the police because it didn’t have the license plate.’
Restrictions
Generic setting	Bare nouns must be used generically; that is, they must denote a maximal set of entities instead of referring to a particular, singular entity.	Goyangyi-nun oo-n-da. Cat-TOP meow-PRES-DEC. ‘The cat meows.’
Superlative context	Bare nouns must be used in a superlative context; that is, they must describe the most unique example of a referred entity.	Sesang-ehseo (*ku) gajang keun nara-nun Russia-i-da. World-LOC (the) most biggest country-TOP Russia-COP-DEC. ‘The biggest country in the world is Russia.’
Referents	Bare nouns can be used for entities that are new to the participants, as well as entities that have already been introduced.	chaek-∅ sa-at-ni? book-∅ buy-PST-INT? 'Did you buy a/the book?'

Definite NPs in Korean are also capable of denoting familiarity. When anaphors are required, both definite nouns and bare nouns can be used to form a grammatical sentence.

Bare noun example:

Eoje	eoddeon	gae-lul	bo-at-da.
Yesterday	a certain	dog-ACC	see-PST-DEC

'I saw a dog yesterday.'

Familiarity as a concept is largely contextual. Here, the speaker presents the entity of the dog as something previously unfamiliar to the listener. There is a unique antecedent in this sentence, so a bare noun can be used.

Definite noun example:

Ku	gae-ka	aju	ddokddok haet-eot-da.
The	dog-NOM	very	smart-PST-DEC.

'The dog was very fluffy.'

(Ku)	gae-ka	aju	ddokddok haet-eot-da.
(That)	dog-NOM	very	smart-PST-DEC.

'(The) dog was very fluffy.'

In these sentences, the speaker refers to the dog as an entity the listener is already familiar with. The sentence is grammatical with a bare noun, but as seen in the second sentence, the demonstrative determiner ku can be used is the context is not enough to give the dog unique identifiability — that is, if there were competing entities for the dog.

In Korean, concepts such as definiteness are able to be encoded within the context of a sentence. Functional elements like overt definite articles (such as an explicit that, or in the case of Korean, a demonstrative such as ku) are not required in the head of a nominal phrase, as we see in English. Korean speakers are able to infer semantic information and identify the entities that bare nouns refer to from within a sentence.

Specificity

Kwon proposes a hypothesis that the specificity interpretation of Korean bare NPs is closely related with their case markers particularly the -nun (TOP), -ga (NOM), -(l)ul (ACC) and the null case -∅. Assuming that D heads are the locus of case features for Korean NPs, he suggests that these case markers which occupy D heads carry not only syntactic features (e.g. case) but also semantic features (e.g. specificity). According to Kwon, topic-marked NPs are always counted as specific, nominative/accusative-marked NPs are interpreted either specific or non-specific depending on the discourse context and lastly, NPs with null case marker always trigger non-specific reading. 

The different semantic strengths of specificity is also expected in Korean depending on which case marker is attached to the NP. Below examples are modified from (8) in the study of Kwon.

(1)  The topic marker –(n)un

Han-sikan	dongan	Lucy-ka	chaek-un	ilk -eot -da. (Specific only)
One-hour	for	Lucy-NOM	book-TOP	read -PST -DEC

‘Lucy read the book for an hour (but not others).’

-nun can be treated as a strong marker for specificity. The interpretation always triggers specific reading.

(2)  The accusative marker –(l)ul & the nominative marker –i or -ka

(a)

Han-sikan	dongan	Lucy-ka	chaek-ul	ilk -eot -da. (Specific or non-specific)
One-hour	for	Lucy-NOM	book-ACC	read -PST -DEC

‘Lucy read a/the book for an hour.’

(b)

Gae-ga	ddokddok-ha-da. (Specific or non-specific)
Dog-NOM	smart-be-DEC

‘A/the dog is smart.’

These two case markers are a neutral marker as they create either a non-specific reading or a specific reading depending on the discourse context.

(3)  The null (phonologically unrealized) case marker -ø

Han-sikan	dongan	Lucy-ka	chaek-ø	ilk -eot -da. (Non-specific only)
One-hour	for	Lucy-NOM	book-ø	read -PST -DEC

‘Lucy read a book for an hour.’

The null case marker is the weakest marker since it always triggers non-specific reading.

To summarize, the specificity interpretation of Korean bare NPs correlates to their morphological endings, particularly the case markers at the D-heads. The difference between specific and non-specific readings on NPs depends on the semantic strength of the nominal markers, which possibly yields strong, neutral and weak specificity.