Brainscape's Knowledge GenomeTM


Top Flashcards (Certified)
All Flashcards

Linguistics > Semantics > Flashcards

Semantics Flashcards

1
Q

Syntax / Semantics Interface
Principle of compositionality

A

The meaning of a phrase / sentence is determined by:
1. the meaning of its parts, and
2. the way they are put together in the syntax

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Role Mapping

A

Meanings are predictable due to the mapping of grammatical roles to semantic roles.

eg: The dog bit Rover.
* the subject of bit is the agent
* the object of bit is the theme

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Productivity

A

Hockett (1960): ‘the capacity to say things that have never been said or heard before and yet to be understood by other speakers of the language’
​
We did not rote-learn all conceivable utterances. But we can generate new utterances, with predictable meanings, by combining familiar parts of utterances in novel ways.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Components of Grammatical Theory

A
  1. Statements about what grammatical information is stored.
    ​
  2. Statements about how you can recursively combine those stored elements into more complex pieces.
    ​
    Learning point: Grammars generate infinitely many sentences from a finite amount of stored information.
    ​
    Goal in semantics is therefore to build a theory of meaning with two components just like the above 1 & 2.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Components of semantic theory

A
  1. Lexical semantics: what information about lexical meaning do we store in our lexicon?
  2. Compositional semantics: how do we predict the meaning of complex expressions from their parts and the way they are combined?

Lexical items are the minimal meaning bearing units. Further complex meanings are created on the basis of them.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Sentence structure to meaning

A

What do the individual words mean?

What structural relationships hold between the items

What principles govern the mapping from smaller pieces of
meaning to larger ones?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Verbs and Arguments

A

1 Argument: Intransitive, 1-place predicate

2 Arguments: Transitive, 2-place predicate

3 Arguments: Ditransitive, 3-place predicate

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Properties of Verbs

A

Verbs force us to consider relations between multiple items.

Relations can be dynamic (eg Mary ate the hot dog) or static (John knows Mary).
* Dynamic: describes events
* Static: describes states

Verb arguments are distinguished by how they participate in these events and states.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Participant Roles / Thematic Relationships

Hinges on prototype concepts that have fuzzy boundaries

A

Participant roles: Different event descriptions often share semantically similar ways that arguments can participate

Examples:
1. Agent: An individual who intentionally initiates the event.
2. Cause: An initiator that may not be acting intentionally but brings about a result.
3. Theme: A participant that (a) undergoes a ‘change of state’ due to an event, or (b) is the target of an emotion.
4. Instrument: A means with which an agent carries out an event
5. Experiencer: an individual in whom an object or event induces some mental state (typically named by the verb).
6. Source / Goal: the initial / final location of the theme
7. Location: the place an event occurs

NB: Participant roles =/= Grammatical roles
* there is a significant subject / non-subject variability

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Agent / Cause / Instrument

A

All of Agent, Cause and Instrument concern things that initiate an event.

  1. Is the thing intentional/conscious in bringing about the event?
    * if YES ==> Agent
    * if NO ==> Q2
  2. Is the thing an instrument used to carry out the event?
    * if YES ==> Instrument
    * if NO ==> Cause
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Argument structure in lexical semantics

A

The lexical entries of items seem to store information about
* the category of the verb
* the arguments it must take

eg: devour
devour: EVENT, PRED = devour, AGENT = x, THEME = y

NB
* information about syntax is not encoded
* usually needs to refer to some kind of external, encyclopaedic knowledge

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Semantic Composition

A

With a 1-place predicate, there is only one open argument slot, which is where the subject must go.

eg: John snores.
snore: EVENT, pred = SNORE, THEME = x
John: j
John snores: EVENT, pred = SNORE, THEME = j

We call something like snore (with empty argument slots) an unsaturated predicate.
John snores (no empty argument slots) is saturated.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Mapping Arguments to Participant Roles

A

The are certain regularities in how participant roles relate to syntactic structure.
* agents / causes are expressed as subjects
* themes are objects unless there are no agents / causes

Hierarchy of grammatical functions map to hierarchy of participant roles.
1. Syntactic: Subject > Direct Object > Indirect Object > Oblique
2. Semantic: Agent / cause > Experiencer > Theme > Goal / source etc

The higher you are on one hierarchy, the higher you will be on the other. (Map semantic roles to the highest available slot in the sytactic hierarchy)

eg: I gave you a book.
I: Subject | Agent
you: Indirect Object | Goal
book: Direct Object | Theme

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Case study: Fear vs Frighten

A

Fear and frighten are very similar, but seem to have their arguments reversed.
* If X fears Y, Y frightens X.

fear: STATE, PRED = fear, EXPERIENCER = x, theme = y
frightens: EVENT, PRED = frighten, CAUSE = x, EXPERIENCER = y

Cause > Experiencer > Theme
* Therefore we expect the experiencer to be a subject in fear, but the object in frightens due to our previous mapping rules.

The argument structure is not actually “reversed”; They actually differ in the nature of participant roles they encode.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Case Study: Passvitisation

A

Problem: active-passive pairs sometimes look like the participant roles are reversed!

eg:
The postman bit Rover (Agent … Theme).
Rover was bitten by the postman (Theme … Agent).

Solution: Passivitisation is a lexical operation that affects a predicate’s participant roles.
* The highest participant role is suppressed
* The next-highest participant role is realised as a subject
* bite: event, pred = bite′, agent = x, theme = y
* bitten: event, pred = bite′, theme = y (the agent / cause becomes an adjunct)

Lexical entries are “silent” about adjuncts / optional modifiers, hence mapping rules don’t apply.

Learning Point: Only a verb’s arguments must observe the mapping rules (cf: Arguments are obligatory)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Learning Points

A

A verb’s meaning gives us a set of participant roles. (Lexically encoded)
1. Can be manipulated by lexical operation
2. Participant roles are projected into syntactic structure
* The sentence has to have the right number of arguments.
* The syntactic position and mapping rules determines how
constituents get associated with participant roles.

Therefore, syntactic structure and lexical semantics determine the interpretation of the sentence (cf: Principle of Compositionality)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Literal Meaning (Semantics)

A

Conventional property of language, independent of use.
* can be understood by any speak of the language, without detailed knowledge of context and speaker

Key metaphor: Language as a conduit (encoding / transmission / decoding)

Deals with sentences: Declarative, Interrogative, Imperative

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

Contextual Meaning (Pragmatics)

A

Key metaphor: Language as social action

Requires integrating knowledge of context and social cognition

Speakers assume active participation of listener: joint action / audience design

Deals with speech acts and beyond: assertions, questions, commands, promises, hints, implicatures

Examples of context dependent meanings:
* definite articles and demonstratives

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

Implicatures (Pragmatics)

A

Additions to meaning generated by counterfactual reasoning about what speaker would have said in various contexts.

Examples
* informativity (quantity)
* truth (quality)
* relevance
* politeness

20
Q

Indirect Speech Acts (Pragmatics)

A

When a sentence type is used to perform a speech act that is not customarily associated with that sentence type
* Arguably, most speech acts are indirect
* Hearers may however be not obliged to consider the literal interpretation first
* As such, syntactic cues are just the starting point for reasoning

21
Q

Presuppositions (Pragmatics)

A
  • Assumptions that are necessary in order for an utterance to make sense
  • Communicatively useful; speakers can exploit them to convey additional information without having to explicitly state the information.

Presuppositions are unified by their behaviour under negation and questioning, which distinguishes them systematically from straightforward entailments.

22
Q

Focus (Pragmatics)

A

The way in which speakers highlight specific information in utterances
* usually done through variations in cadence and pitch for English

23
Q

Importance of Literal Meaning

A

Pragmatics generally take literal meaning as a starting point.
* meaning is too dependent on syntax and context
* literal meaning cannot be fully resolved without knowledge of context as well

Semantics determined systematically by
* meaning of lexical items
* syntax (how they are put together)
* contextual information (eg filling demonstratives)

24
Q

Lexical Meanings

In your head stored information

A

Extension: The real world referent that the word ‘picks out’

Intension: What language users know about the word that would, with sufficient information, allow them to work out the extension.
* roughly, the concept that is associated with the word

25
Return to compositionality
Interpreting a "new" sentence: * meaning of lexical items * systematic principles for combining them into a sentence meaning (following the syntactic structure of the sentence) Recall: Principle of Compositionality The meaning of a sentence is determined by the meanings of its parts, and the way they are put together.
26
Key types of evidence in Semantics
1. Would sentence S be true in the following scenario? (information about TRUTH CONDITITIONS) 2. Does sentence S entail sentence T? * if we assume S is true, do we have to count T as true also? * same for 'contradict', 'mean the same as', 'are they compatible' Traditionally, investigated using intuition, but more recently has been subject to corpus analysis and online / offline experiments.
27
Sense Relations
1. Synonymy (same meaning // equivalence of sense) 2. Entailment / Inclusion / Part-whole relations 3. Compatibility (both can be true while being independent) 4. Contradiction / exclusion (S and ¬S)
28
Entailment / Inclusion / Hyponymy
Hyponymy is about the different subcategories of a word's denotation * deals with denotional subset A sentence containing a hyponym entails the corresponding sentence in which the hyponym has been replaced by its superordinate. Hierarchies of Hyponyms - As we move up the hierarchies of hyponyms, the senses of the words become **less specific** and the denotations become **larger and more general** - At lower levels the senses are more **detailed** and the words denote **narrower ranges of things** - We can try to capture the meaning of each hyponym as the meaning of its **immediate superordinate with a modifier** (eg house is a building for living in) ○ This captures the key idea that a hyponym is a special case of its superordinate
29
Contradiction
A and B are contradictory if both cannot be simultaneously true. Subsets of contadiction: 1. Complementarity * Pairs of words that divide a class of objects into two non-overlapping sets * Complement of the things that are not described by one term are described by the other 2. Antonymy * Weaker than complementarity, used to describe a specific kind of "oppositeness" * Pairs of antonyms tap into meanings that are at opposite extremes, but unlike complementaries, leave gaps in the middle. 3. Converse * Adjectival antonym pairs entails a sense relation between their comparative forms as well.
30
Synonymy
To understand synonymy, one must distinguish **sense** and **reference**. Synonymy compares the **sense**, while the references can be different
31
Lexical Meaning
Linguistic meanings relate to **mental constructs** There is a **systematic** way in which we understand these words, and how they are affected by context. Lexical meaning, like the rest of grammar, is therefore part of our **implicit knowledge of language**
32
Concepts
Idea -> Mental Object -> Concept Concepts need not be mental representations of individual parts of the world. The concept of a word must therefore be a sort of "essence" (or **prototype**) of the word.
33
Aristotelian (=classical) theory of concepts
A set of properties that characterises **all and only** instances of the concept (assumes there are enough conditions to uniquely categories). Limitations: * mainly applied to things like geometry with clear and discrete categories * **concepts cannot be captured by necessary and sufficient conditions** * does not properly deal with exceptions
34
Boundaries between concepts
case study: Labov's cups and bowls Addition of mashed potato / flowers affected perception of cup / bowl / vase. Learning point: Concepts have vague boundaries that require external concepts to better define.
35
Prototypes
cf: Rosch study Prototype theory: word meanings are organised (at least in part) around prototypes rather than obligatory properties (does not entail specific properties, but entails enough of the prototypical properties to be termed a certain noun) Prototypes are: * culturally dependent * but systematic
36
Relational Concepts
Concepts are defined **with reference to each other** Variation depends on how well an entity represents a concept (or having a greater number of similarities with the protoype). eg: Colour Terms Light + Dark > Red > Yellow > Green
37
Concepts and reasoning
Concepts are used by reasoning with them. We can reason more **productively with concepts** because we know that certain relations hold between them. 1. Hyperonymy (More generic concept than) / Hyponym (more specific concept than) 2. Meronymy (a part of a whole) 3. Antonymy (opposites) 4. Synonymy (A = B and vice versa)
38
Hierarchical Relations
Relations of **hyponymy and hyperonymy** are **transitive** (mathematically) Concepts can be organised into a **hierarchy** * even though concepts usually do not have clear definitions, this does not affect our semantic reasoning
39
Sense (intension) and reference (extension)
An expressions sense / intension (the concept / property it names) from its reference / extension (the real world objects that **realise** the concept or property). Words like unicorn have a clear intension, but an **empty extension** General Rule: Intension + **Facts of Time and Space** = Extension
40
Real World Referencing
Key distinctions > individuals and properties > referring expressions as names for individuals (eg: the table in front of me...) > Predicates as names for properties (eg: is hard) > Intensions vs extensions In general, Referring expression + predicate = sentence.
41
Semantic Model
1. A set of **individuals**, or specific objects. 2. Some **properties** individuals are necessarily one-offs, properties can apply to several (or no) individuals **Individuals have properties**
42
Predicates: Syntax-Semantics Interface
Predicates are a syntactic category * they denote **properties** Properties are a **semantic** category * members of different syntactic categories can denote properties
42
Referring Expression
RE: any noun phrase, or surrogate for a noun phrase, whose function in discourse is to identify some individual object > eg Proper Names RE **felicity conditions** > picks out a specific individual in the speaker's mental model > (speaker believes that) it picks out a specific individual in the hearer's mental model > (speaker believes that) those two individuals refer to 'the same individual'
43
Intensional and extensional predicates
Certain predicates are **extensional** in that they require us to *interpret tem as describing objects in the real world* > 'looks at' the extension of its object Others are **intensional** in that they concern beliefs, hypothetical sitatuations. > 'looks at' the intension of its object eg: Joe Biden **met** Barack Obama (ext.) Joe Biden **discussed** James Bond (int.) **Test**: If replacing the object NP with another NP with the same extension but different intension can change the sentence meaning, the predicate is **intensional**. Otherwise it is **extensional**
44
Context
A provisional model of context can be given as **situation + time** Intension + context ==> extension > mathematically, the intension is a function that maps from context to extension > given a context an **intension determines an extension**
44
Dynamic nature of extensions
The extension of an expression often depends on **context** and **how the world is** Two expressions can have the same expression in the here-and-now, but different extensions in other contexts. **Extensions shift with contexts; intensions don't**

Linguistics (7 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2025 Bold Learning Solutions. Terms and Conditions