THE MASTERY CLUB · VOCABULARY SERIES · SEMANTIC ROLES · THEMATIC ROLES · THETA ROLES · WHO DID WHAT TO WHOM

Vocabulary | Semantic Roles — How Agent, Patient, Experiencer, Instrument, Source and Goal Structure Meaning

Semantic roles—also called thematic roles, thematic relations, participant roles or, in some grammatical traditions, theta roles—describe the relationship that a noun phrase or participant bears to the event, state or relation expressed by a predicate. Labels such as Agent, Patient, Theme, Experiencer, Stimulus, Instrument, Recipient, Beneficiary, Source, Goal, Path and Location answer questions such as: Who intentionally acted? What changed? Who experienced the emotion? What moved? What tool was used? Where did the movement begin and end? Search questions such as “What are semantic roles?”, “What is the difference between Agent and Patient?”, “Is the subject always the Agent?”, “What are thematic roles?”, “What is Theme vs Patient?”, “What is an Experiencer?” and “How do semantic roles connect meaning to grammar?” all point toward the same central problem: sentences organise participants by meaning as well as by syntax.

Semantic role is not the same thing as grammatical function. In Maya broke the glass, Maya is both grammatical subject and semantic Agent, while the glass is object and Patient. But in The glass was broken by Maya, the glass becomes grammatical subject while remaining the Patient, and Maya remains the Agent inside a by-phrase. In The glass broke, the glass is again subject but is still Patient- or Theme-like rather than Agentive. In The key opened the lock, the subject is an Instrument or Cause-like participant, not an intentional Agent. Learning semantic roles therefore prevents the common mistake of treating “subject = doer” and “object = receiver of action” as universal definitions.

This article is the canonical Semantic Roles specialist inside Vocabulary — The Mastery Club apex. It is a completely new article and does not rewrite existing eduKate pages. The companion Vocabulary | Frame Semantics owns scene-specific roles such as Buyer, Seller, Goods and Money; this page owns the broader cross-predicate abstractions such as Agent, Patient, Experiencer, Instrument, Source and Goal. Existing pages on What Is Semantics?, complements/adjuncts, grammatical case, verb patterns and argument structure remain protected. The job here is specific: explain participant roles, distinguish them from subjects and objects, show how roles interact with verb meaning and sentence alternations, and make the system useful for vocabulary, reading, grammar and writing.

The 60-Second Answer: What Are Semantic Roles?

Semantic roles classify participants according to how they take part in an event or state. In Maya opened the door with a key, Maya is Agent because she intentionally initiates the action; the door is Patient because it undergoes a change of state; the key is Instrument because it is used to carry out the action. In Maya fears spiders, Maya is Experiencer and spiders are Stimulus. In Arun sent the parcel from Singapore to Tokyo, Arun is Agent or Sender, the parcel is Theme, Singapore is Source and Tokyo is Goal.

There is no single universally agreed list of semantic roles. Linguistic traditions vary in labels, granularity and theoretical status. Oxford’s computational-linguistics glossary notes that role names can overlap—for example Theme is sometimes used where other accounts use Patient—and Cambridge discussions of argument realisation explicitly review the limitations of fixed role inventories. David Dowty’s influential proto-role theory replaces a rigid list for some purposes with two cluster concepts, Proto-Agent and Proto-Patient, defined by bundles of entailments rather than indivisible categories.

For teaching, the best approach is pragmatic: use a small, explicit role inventory to clarify who participates and how, while recognising that some examples sit between labels. The role system is a map of recurring semantic patterns, not a claim that every argument in every language belongs to one perfectly bounded box.

Part I — The Core Role Inventory

1. Agent

The participant that intentionally initiates or controls an action. In Maya opened the window, Maya is Agent. Typical Agent-like properties include volition, awareness, causation and independent movement.

2. Patient

The participant that undergoes a change of state or is strongly affected. In Maya broke the glass, the glass is Patient because its state changes.

3. Theme

A participant that moves, is located, or is otherwise centrally involved without necessarily undergoing internal change. In Maya moved the box to the table, the box is Theme. Terminology varies: some traditions use Theme more broadly and overlap it with Patient.

4. Experiencer

The sentient participant that feels, perceives or mentally experiences a state. In Maya fears spiders, Maya is Experiencer. In Maya heard the bell, Maya is Experiencer.

5. Stimulus

The entity or situation that triggers an experience or perception. In Spiders frighten Maya, spiders are Stimulus and Maya is Experiencer.

6. Instrument

The entity used to perform an action. In Maya opened the door with a key, the key is Instrument.

7. Recipient

The participant that receives a Theme in a transfer. In Maya gave Arun the book, Arun is Recipient.

8. Beneficiary

The participant for whose benefit an event is carried out. In Maya baked Arun a cake, Arun can be Beneficiary even if the cake has not yet been transferred.

9. Source

The starting point of motion, transfer or a relation. In The train travelled from Paris, Paris is Source.

10. Goal

The endpoint or destination of motion or transfer. In The train travelled to Berlin, Berlin is Goal.

11. Path

The route along which movement occurs. In The train travelled through the tunnel, the tunnel can realise Path.

12. Location

A place where an entity or event is situated. In The book is on the table, the table/location phrase expresses Location.

13. Cause

A non-agentive participant or event that brings about another event. In The storm broke the window, the storm is Cause-like rather than a volitional Agent.

14. Possessor

The participant that owns or has control/association with an entity. In Maya has a bicycle, Maya is Possessor in a possession relation.

15. Attribute / Value

In some semantic analyses, states and measurements involve an Entity and an Attribute/Value: The box weighs ten kilograms. Role labels vary considerably here.

Part II — Semantic Roles Are Not Grammatical Functions

16. Subject can be Agent

Maya opened the door. Subject = Maya; role = Agent.

17. Subject can be Patient

The door was opened by Maya. Subject = the door; role = Patient.

18. Subject can be Theme

The ball rolled down the hill. Subject = ball; role = Theme.

19. Subject can be Experiencer

Maya likes jazz. Subject = Maya; role = Experiencer.

20. Subject can be Stimulus/Cause-like

The noise annoyed Maya. Subject = the noise; role = Stimulus/Cause-like participant.

21. Subject can be Instrument

This key opens the door. Subject = key; role can be analysed as Instrument/Cause-like depending theory.

The syntactic category “subject” tells us how an expression behaves grammatically. A semantic role tells us how the participant relates to the event. They often correlate, but they are not interchangeable concepts.

Part III — Passive Voice Proves Subject Is Not Agent

22. Active

Maya broke the glass. Maya is Agent and grammatical subject. The glass is Patient and object.

23. Passive

The glass was broken by Maya. The glass becomes grammatical subject but remains Patient. Maya remains Agent even though it is expressed in a by-phrase.

24. Agentless passive

The glass was broken. The Agent is omitted. The event is still understood as involving a cause or agentive source, but grammar does not require that participant to be stated.

25. Why this matters

If learners define subject as “the doer,” passive clauses become confusing. Semantic roles give a stable meaning layer while syntax changes.

Part IV — Causative and Inchoative Alternations

26. Causative

Maya broke the glass. Agent/Cause = Maya; Patient = glass.

27. Inchoative

The glass broke. No external Agent is expressed. The glass is Patient/Theme subject and undergoes change.

28. Melt

The heat melted the ice versus The ice melted. The ice retains the change-of-state role even though its grammatical function changes.

29. Open

Maya opened the door versus The door opened. The door is the entity whose state changes.

30. Dry

The sun dried the clothes versus The clothes dried. Cause and Patient roles remain conceptually distinct from subject/object.

Part V — Agent vs Cause

Not every causer is an Agent. Intentionality and control are central diagnostics in many role traditions.

31. Volitional Agent

Maya deliberately broke the branch. Maya intends and controls the action.

32. Natural Cause

The wind broke the branch. The wind causes the event but is not normally treated as a conscious Agent.

33. Event Cause

The collision broke the axle. The collision is a Cause-like event.

34. Instrument as subject

The key opened the door. The key is not a volitional Agent, yet English allows it as subject. Some theories analyse it as Instrument, Cause or an Agent-like proto-role depending the predicate.

35. Why labels sometimes differ

These borderline cases are one reason semantic-role inventories vary. A useful analysis states the properties involved rather than pretending every subject must fit one rigid category.

Part VI — Patient vs Theme

36. Strongly affected Patient

The chef sliced the bread. The bread undergoes an internal physical change.

37. Moving Theme

Maya carried the box upstairs. The box changes location but need not change internally.

38. Located Theme

The box remained on the shelf. The box is a located entity rather than an affected Patient.

39. Terminological overlap

Some textbooks use Theme broadly enough to include many Patient-like arguments. Others reserve Patient for affected entities and Theme for moved/located entities. The important teaching distinction is affectedness versus participation/motion, not winning a terminology dispute.

Part VII — Experiencer and Stimulus

40. Experiencer-subject verb

Maya fears spiders. Maya is Experiencer; spiders are Stimulus.

41. Stimulus-subject verb

Spiders frighten Maya. Spiders are Stimulus/Cause-like; Maya remains Experiencer even though she is object.

42. Perception

Maya heard the bell. Maya is Experiencer; bell/sound is Stimulus or Theme depending framework.

43. Mental state

Arun knows the answer. Arun is Experiencer/Knower-like; the answer is Content/Theme-like.

44. Emotion adjectives

Maya is afraid of spiders. Maya remains Experiencer despite adjective syntax. Semantic roles can span grammatical categories.

Part VIII — Recipient and Beneficiary

45. Recipient

Maya gave Arun the book. Arun receives the Theme and is Recipient.

46. Beneficiary

Maya baked Arun a cake. Arun benefits from the event. He may become Recipient later, but the sentence primarily encodes Beneficiary.

47. Recipient can be Goal-like

Transfer can be conceptualised as Theme moving toward a Recipient. This explains why role systems sometimes connect Recipient with Goal while still distinguishing human transfer from spatial destination.

48. Maleficiary

Some analyses recognise a participant adversely affected by an event. Many teaching systems simply treat this as Beneficiary-like affectedness with negative outcome rather than a separate core role.

Part IX — Source, Goal and Path

49. Source

The train left Paris. Paris is Source: the starting point.

50. Goal

The train arrived in Berlin. Berlin is Goal: the endpoint.

51. Path

The train travelled through the tunnel. The tunnel expresses Path.

52. Theme in motion

The parcel moved from Singapore through Bangkok to Tokyo. Parcel = Theme; Singapore = Source; Bangkok = Path/Way-point; Tokyo = Goal.

53. Agentive mover

Maya walked from home to school. Maya is both the intentional controller of walking and the entity that moves. Different theories may emphasise Agent, Theme or proto-role properties. This is a useful example of role overlap.

Part X — Instrument, Means and Manner

54. Instrument

Maya cut the rope with a knife. Knife is Instrument.

55. Means

Maya travelled by train. The train can be treated as Means/Vehicle rather than a simple physical Instrument.

56. Manner

Maya spoke softly. Softly expresses Manner, not a participant role in the same sense as Agent or Patient. Some broad semantic-role inventories include Manner-like adjunct roles.

57. Cause vs Instrument

The hammer broke the glass can profile the instrument as subject, while Maya broke the glass with a hammer separates Agent and Instrument. The event can be linguistically framed at different causal levels.

Part XI — Location, Source, Goal and Place Expressions

58. Static Location

The keys are in the drawer. Drawer phrase expresses Location.

59. Event Location

The concert took place in the hall. Hall locates the event rather than a moving Theme.

60. Abstract Location

Technical and metaphorical language can extend spatial roles: in the category, within the system, from theory to practice. Not every spatial-looking phrase is a literal physical location.

Part XII — Possessor and Possession Relations

61. Possessor

Maya owns a bicycle. Maya is Possessor; bicycle is Possessed/Theme-like entity.

62. Have

Maya has a bicycle. English uses a broad possession relation that can express ownership, kinship, attributes or states.

63. Belong

The bicycle belongs to Maya. The same possession relation is profiled from the possessed entity.

64. Frame-specific refinement

Frame semantics might use Owner and Possession as frame-specific roles. General semantic-role theory may simply use Possessor and Theme. The two levels of description are complementary.

Part XIII — Semantic Role Laboratory: 60 Sentences

1. Maya opened the door.

Roles: Maya = Agent; door = Patient. Why: Intentional action changes door state.

2. The door opened.

Roles: door = Patient/Theme. Why: Affected entity becomes subject without Agent.

3. The wind opened the door.

Roles: wind = Cause; door = Patient. Why: Non-volitional causer.

4. The key opened the door.

Roles: key = Instrument/Cause-like; door = Patient. Why: Instrument can surface as subject.

5. Maya cut the bread with a knife.

Roles: Maya = Agent; bread = Patient; knife = Instrument. Why: Agent and Instrument overtly separated.

6. The knife cut the bread easily.

Roles: knife = Instrument/Cause-like subject; bread = Patient. Why: Subject need not be Agent.

7. The chef melted the butter.

Roles: chef = Agent; butter = Patient. Why: Causative change of state.

8. The butter melted.

Roles: butter = Patient/Theme. Why: Inchoative alternation.

9. The sun dried the clothes.

Roles: sun = Cause; clothes = Patient. Why: Natural cause.

10. The clothes dried in the sun.

Roles: clothes = Patient; sun phrase = Location/Cause context. Why: Affected subject.

11. Maya fears spiders.

Roles: Maya = Experiencer; spiders = Stimulus. Why: Emotion state.

12. Spiders frighten Maya.

Roles: spiders = Stimulus/Cause; Maya = Experiencer. Why: Same conceptual relation, different lexical perspective.

13. Maya heard the bell.

Roles: Maya = Experiencer; bell = Stimulus/Theme. Why: Perception.

14. The bell startled Maya.

Roles: bell = Stimulus/Cause; Maya = Experiencer. Why: Stimulus subject.

15. Arun knows the answer.

Roles: Arun = Experiencer/Knower; answer = Content/Theme. Why: Mental state.

16. The answer surprised Arun.

Roles: answer = Stimulus; Arun = Experiencer. Why: Emotion/cognition change.

17. Maya gave Arun a book.

Roles: Maya = Agent/Giver; Arun = Recipient; book = Theme. Why: Transfer.

18. Arun received a book from Maya.

Roles: Arun = Recipient; book = Theme; Maya = Source/Giver. Why: Same transfer from recipient perspective.

19. Maya baked Arun a cake.

Roles: Maya = Agent; Arun = Beneficiary; cake = Created Theme. Why: Benefactive construction.

20. Maya sent the parcel to Tokyo.

Roles: Maya = Agent/Sender; parcel = Theme; Tokyo = Goal. Why: Caused motion/transfer.

21. The parcel travelled from Singapore to Tokyo.

Roles: parcel = Theme; Singapore = Source; Tokyo = Goal. Why: Theme moves without overt Agent.

22. Maya walked from home to school.

Roles: Maya = Agent-like mover/Theme; home = Source; school = Goal. Why: Role overlap in self-motion.

23. The bird flew across the river.

Roles: bird = Agentive Theme; river = Path. Why: Mover both controls and undergoes motion.

24. The leaf drifted across the pond.

Roles: leaf = Theme; pond = Path/Location. Why: Non-agentive movement.

25. Maya travelled by train.

Roles: Maya = Agentive Theme; train = Means/Vehicle. Why: Means differs from Instrument.

26. Maya wrote the note with a pencil.

Roles: Maya = Agent; note = Created Theme; pencil = Instrument. Why: Creation plus instrument.

27. The pencil wrote smoothly.

Roles: pencil = Instrument subject. Why: Metonymic/instrumental subject construction.

28. The storm destroyed the roof.

Roles: storm = Cause; roof = Patient. Why: Natural Cause affects Patient.

29. The explosion shattered the windows.

Roles: explosion = Cause; windows = Patient. Why: Event Cause.

30. Maya pushed the trolley.

Roles: Maya = Agent; trolley = Theme/Patient-like. Why: Motion without necessary internal change.

31. Maya crushed the can.

Roles: Maya = Agent; can = Patient. Why: Strong affectedness.

32. Maya moved the chair.

Roles: Maya = Agent; chair = Theme. Why: Location change.

33. The chair moved.

Roles: chair = Theme. Why: Inchoative/motion subject.

34. The book lies on the table.

Roles: book = Theme; table = Location. Why: Static location.

35. The children played in the park.

Roles: children = Agents/Actors; park = Location. Why: Location adjunct.

36. The meeting happened in the hall.

Roles: meeting event = Theme-like eventuality; hall = Location. Why: Event location.

37. Maya put the cup on the shelf.

Roles: Maya = Agent; cup = Theme; shelf = Goal/Location. Why: Put strongly selects a location argument.

38. Maya removed the cup from the shelf.

Roles: Maya = Agent; cup = Theme; shelf = Source. Why: Removal reverses spatial relation.

39. The teacher explained the rule to the class.

Roles: teacher = Agent/Speaker; rule = Content; class = Recipient/Addressee. Why: Communication.

40. The class heard the explanation.

Roles: class = Experiencer/Recipient; explanation = Stimulus/Content. Why: Reception perspective.

41. Maya told Arun the news.

Roles: Maya = Agent/Speaker; Arun = Recipient/Addressee; news = Content. Why: Communication transfer.

42. Arun learned the news from Maya.

Roles: Arun = Experiencer/Learner; news = Content; Maya = Source. Why: Knowledge acquisition.

43. The report informed the public.

Roles: report = Stimulus/Information Source; public = Recipient/Experiencer. Why: Inanimate subject.

44. The headline misled readers.

Roles: headline = Stimulus/Cause; readers = Experiencers. Why: Cognitive effect.

45. Maya owns the bicycle.

Roles: Maya = Possessor; bicycle = Possessed/Theme. Why: Possession.

46. The bicycle belongs to Maya.

Roles: bicycle = Theme/Possessed; Maya = Possessor/Goal-like relation. Why: Possession from item perspective.

47. Maya owes Arun ten dollars.

Roles: Maya = Debtor; Arun = Creditor/Recipient-like; ten dollars = Theme/Amount. Why: Debt relation; frame-specific roles refine general roles.

48. The bank lent Maya money.

Roles: bank = Agent/Lender; Maya = Recipient/Borrower; money = Theme. Why: Loan transfer.

49. Maya borrowed money from the bank.

Roles: Maya = Recipient/Borrower; money = Theme; bank = Source/Lender. Why: Perspective switch.

50. The doctor prescribed Maya antibiotics.

Roles: doctor = Agent/Prescriber; Maya = Recipient/Patient; antibiotics = Theme. Why: Institutional transfer/instruction.

51. Maya took the medicine.

Roles: Maya = Agent/Patient-like consumer; medicine = Theme. Why: Same healthcare frame, different participant action.

52. The medication relieved the pain.

Roles: medication = Cause/Instrument-like; pain = Patient/affected state. Why: Inanimate cause.

53. The judge sentenced the defendant.

Roles: judge = Agent/Authority; defendant = Patient/Recipient of institutional action. Why: Legal frame.

54. The defendant received a sentence.

Roles: defendant = Recipient/Patient-like; sentence = Theme. Why: Perspective switch.

55. The witness gave evidence.

Roles: witness = Agent/Source; evidence = Theme. Why: Communication/legal frame.

56. The evidence convinced the jury.

Roles: evidence = Stimulus/Cause; jury = Experiencer. Why: Cognitive causation.

57. The company hired Maya.

Roles: company = Agent/Employer; Maya = Patient/Employee candidate. Why: Institutional change-of-state relation.

58. Maya joined the company.

Roles: Maya = Agentive Theme; company = Goal/Institution. Why: Same employment system, self-motion metaphor.

59. The committee approved the proposal.

Roles: committee = Agent/Evaluator; proposal = Theme/Target. Why: Decision frame.

60. The proposal satisfied the criteria.

Roles: proposal = Theme/Target; criteria = Standard-like role. Why: Relational evaluation; no intentional Agent required.

Part XIV — Role Labels Can Overlap

Some participants satisfy properties associated with more than one role. A person who deliberately walks to school is Agent-like because they act intentionally and Theme-like because they undergo motion. A medication can be Instrument-like because it is used in treatment and Cause-like because it produces physiological change. Theoretical approaches handle these cases differently.

65. Role overlap is evidence, not failure

If two labels seem plausible, describe the relevant properties: intentionality, movement, affectedness, sentience, causation and change of state.

66. Frame-specific labels can be more precise

In a healthcare frame, Patient as an institutional role can coexist with semantic Patient/Experiencer distinctions. Capitalisation and context can help avoid confusion.

67. Teaching should prioritise contrastive value

Use labels when they explain a real difference: Experiencer versus Stimulus, Recipient versus Beneficiary, Source versus Goal, Agent versus Cause.

Part XV — Semantic Roles and Argument Structure

Predicates differ in how many participants they conventionally select and what semantic relations those participants bear. A semantic-role list or theta-grid is one way of representing that information.

Part XVI — Dative Alternation: Same Roles, Different Syntax

68. Double-object

Maya gave Arun the book. Maya = Agent/Giver; Arun = Recipient; book = Theme.

69. To-dative

Maya gave the book to Arun. The semantic roles are essentially the same even though the grammatical structure changes.

70. Lexical restrictions

Donate shares a transfer meaning but normally prefers donate the book to the library, not *donate the library the book*. Semantic similarity does not guarantee identical argument realisation.

71. Information structure

Choice between constructions can interact with discourse factors such as givenness, weight and focus. Roles stay semantic while syntax manages packaging.

Part XVII — Locative Alternation

72. Theme-locative

Maya loaded hay onto the wagon. Hay = Theme; wagon = Goal/Location.

73. Location-with

Maya loaded the wagon with hay. Wagon becomes object and is often interpreted as more completely affected.

74. Semantic consequence

Alternations can change construal as well as syntax. Roles illuminate the participants while construction choice adds affectedness or perspective.

Part XVIII — Resultative Constructions

75. Action plus result

Maya painted the door red. Maya = Agent; door = Patient; red = Result State.

76. Created result

Maya hammered the metal flat. Metal is Patient; flat expresses the endpoint state.

77. Why roles help

The sentence contains more than Agent–Patient. Result State is semantically crucial even though it is not another participant noun phrase.

Part XIX — Experiencer Alternations

78. Fear

Maya fears spiders. Experiencer subject, Stimulus object.

79. Frighten

Spiders frighten Maya. Stimulus/Cause subject, Experiencer object.

80. Please

The music pleased Maya. Stimulus subject, Experiencer object.

81. Like

Maya likes the music. Experiencer subject, Stimulus/Theme object.

82. Cross-linguistic caution

Languages can encode experiencers with different grammatical cases or positions. The semantic relation should not be identified with English subject/object patterns.

Part XX — Passive, Voice and Role Preservation

Voice changes how arguments map onto grammatical functions. Semantic roles remain a stable layer for tracking meaning.

Maya wrote the report.

Roles: Maya Agent; report Theme/Created Object. Lesson: Active subject is Agent.

The report was written by Maya.

Roles: report Theme subject; Maya Agent by-phrase. Lesson: Passive changes syntax, not roles.

The report was written.

Roles: report Theme subject; Agent unexpressed. Lesson: Role can be conceptually implicit.

The child was frightened by the noise.

Roles: child Experiencer; noise Stimulus/Cause. Lesson: Passive subject is Experiencer, not Patient in a simple action sense.

The parcel was sent to Tokyo by Maya.

Roles: parcel Theme; Tokyo Goal; Maya Agent. Lesson: Several semantic roles survive passive rearrangement.

Part XXI — Semantic Roles and Grammatical Case

Languages can mark participant roles through case endings, adpositions, word order, agreement or combinations of these. Grammatical case and semantic role interact, but one case does not always equal one role.

83. Nominative

Often marks subjects, which can be Agent, Experiencer, Theme, Patient or other roles.

84. Accusative

Often marks objects, which can be Patient, Theme, Stimulus or other roles.

85. Dative

Frequently associated with Recipient, Experiencer or Beneficiary functions across languages, but systems differ.

86. Instrumental

Some languages morphologically mark Instrument/Means roles that English usually expresses with prepositions such as with or by.

87. Locative/allative/ablative

Case systems can distinguish Location, Goal (“to”) and Source (“from”) directly.

88. Why translation needs roles

When languages mark participants differently, translation should preserve who plays which semantic role rather than copy word order or case mechanically.

Part XXII — Semantic Roles and Prepositions

Part XXIII — Selectional Preferences

89. Agents tend to be capable of causation/control

Verbs such as decide, apologise, promise strongly favour sentient intentional subjects.

90. Experiencers tend to be sentient

Fear, notice, remember normally involve beings capable of experience, though metaphorical usage can stretch these preferences.

91. Instruments fit action affordances

Cut with a knife is semantically ordinary because a knife affords cutting.

92. Selectional violations can signal metaphor

The headline screamed danger gives a nonhuman subject a speech-like action. Readers reinterpret roles metaphorically rather than reject the sentence.

Part XXIV — Semantic Roles and Vocabulary Precision

Role knowledge sharpens near-synonym distinctions because many word differences are differences in participant structure.

borrow / lend

Role distinction: Borrower vs Lender perspective. Word-choice lesson: Same Theme transfer, opposite participant foreground.

teach / learn

Role distinction: Source/Teacher vs Learner perspective. Word-choice lesson: Knowledge-transfer roles.

give / receive

Role distinction: Giver vs Recipient perspective. Word-choice lesson: Converse transfer.

buy / sell

Role distinction: Buyer vs Seller perspective. Word-choice lesson: Commercial roles.

fear / frighten

Role distinction: Experiencer vs Stimulus subject. Word-choice lesson: Emotion roles.

win / beat

Role distinction: Winner + Competition vs Winner + Opponent. Word-choice lesson: Object selection differs.

rob / steal

Role distinction: rob selects Victim/Place; steal selects Stolen Theme. Word-choice lesson: Crime roles determine grammar.

provide / supply

Role distinction: Provider, Recipient, Theme patterns overlap but constructions differ. Word-choice lesson: Valence matters.

explain / tell

Role distinction: Content/Recipient grammatical realisation differs. Word-choice lesson: Semantic role similarity, syntax difference.

load / fill

Role distinction: Theme-Location alternation and affectedness differ. Word-choice lesson: Roles plus construction semantics.

Part XXV — Research Boundary: No Universal Role List

Semantic-role labels have a long history in linguistics, but there is no single universally accepted inventory. Cambridge’s Argument Realization describes semantic role lists—also called case frames or theta-grids—as a widespread way to represent grammatically relevant aspects of verb meaning. Typical labels include Agent, Patient, Theme, Experiencer, Instrument, Source, Location and Goal. The same source also emphasises fundamental limitations of fixed role-list approaches.

Oxford’s Handbook of Computational Linguistics glossary similarly defines a semantic role as the relation a noun phrase bears to a verb and illustrates movement with Agent, Theme, Source, Path, Goal and Manner. It explicitly notes terminological overlap: Theme may be used where another tradition uses Patient, and Source may receive alternative labels. This is why the article treats role labels as useful analytical tools rather than immutable natural kinds.

Cambridge work on subjects and case relations also shows why grammatical function must be kept separate from semantic role. Subjects can express Agents, Experiencers, Instruments, Patients or Themes. Linguists have proposed role hierarchies to capture tendencies in subject selection, but there is little agreement on one precise universal hierarchy.

• Oxford Handbook of Computational Linguistics — semantic role glossary
• Cambridge — Semantic Role Lists in Argument Realization
• Cambridge — Thematic Relations
• Cambridge — Agent and Subject

Part XXVI — Dowty’s Proto-Agent and Proto-Patient

David Dowty’s influential proto-role approach responds to the difficulty of drawing hard boundaries among thematic roles. Rather than assuming every argument belongs to exactly one atomic role, the theory identifies clusters of properties associated with two prototypes: Proto-Agent and Proto-Patient. The same paper was republished online by Cambridge in 2026, reflecting its continuing importance in current role theory.

93. Proto-Agent properties

Typical Proto-Agent properties include volitional involvement, sentience/perception, causing an event or change of state in another participant, movement, and independent existence relative to the event. An argument can show several of these properties without matching every one.

94. Proto-Patient properties

Typical Proto-Patient properties include undergoing change of state, being causally affected, being stationary relative to another participant’s movement, lacking independent existence relative to the event in some cases, and being incrementally affected by the event.

95. Roles can be gradient

An argument may be more Agent-like or Patient-like depending on how many properties it has. This handles cases where rigid labels feel forced.

96. Example: self-motion

Maya ran to school. Maya is volitional, sentient and moving—strong Proto-Agent properties—but is also the entity whose location changes, a Theme-like property. Proto-role analysis allows this mixed profile.

97. Example: instrument subject

The key opened the door. The key causes the change but lacks sentience and volition. It has some Proto-Agent properties but not the full prototypical Agent cluster.

98. Example: experiencer

Maya fears spiders. Maya is sentient and perceives/experiences, but does not cause a change in the Stimulus. The argument is subject-like and Proto-Agent-like in some respects without being a canonical action Agent.

99. Why proto-roles help teaching

When students encounter borderline cases, ask which semantic properties are present rather than forcing one label. The goal is to understand participant structure, not win a classification contest.

Part XXVII — Role Hierarchies

Many theories propose hierarchies such as Agent > Experiencer > Theme/Patient > Goal or related variants to describe preferences in grammatical mapping. Recent 2025 work in Language and Cognition discusses an Agent > Experiencer > (Causer)/Theme prominence pattern in processing. But role hierarchies differ across frameworks and should be treated as tendencies/theoretical tools rather than one settled universal order.

100. Canonical subject mapping

English active clauses often place highly Agent-like participants as subjects.

101. Exceptions are systematic

Experiencer subjects, Patient passive subjects, Theme unaccusative subjects and Instrument subjects demonstrate that subjecthood is grammatically broader than Agenthood.

102. Processing prominence

Listeners may use broad semantic prominence to anticipate sentence structure, but finer role distinctions emerge from the predicate and full clause.

Part XXVIII — Semantic Roles vs Theta Roles

103. Thematic role

A broad semantic label describing participant relation to an event or predicate.

104. Theta role

In generative grammar, theta roles are linked to a predicate’s argument structure and principles governing how arguments receive semantic roles. Usage varies by theory.

105. Theta-grid

A representation of the roles/arguments selected by a predicate. A simple teaching approximation might list give: Agent, Theme, Recipient.

106. Why the terms should not be collapsed carelessly

Semantic-role teaching can remain descriptive without committing to every theoretical claim of theta theory. This article uses thematic/semantic role broadly and introduces theta terminology for readers encountering grammar literature.

Part XXIX — Semantic Roles vs Frame Elements

The companion Frame Semantics article uses specific roles such as Buyer, Seller, Goods and Money. Semantic roles generalise across many frames.

Part XXX — What Semantic Roles Do Not Claim

107. They do not equal grammatical cases

One case marker can realise several semantic roles, and one role can be marked differently across constructions/languages.

108. They do not equal subject/object

Subjects and objects are syntactic functions; roles are semantic relations.

109. They do not always have sharp boundaries

Theme/Patient and Agent/Cause/Instrument can overlap depending theory and example.

110. They do not replace frame-specific analysis

Buyer/Seller or Doctor/Patient may be more informative than Agent/Recipient in a particular domain.

111. They do not replace lexical meaning

Knowing that a verb takes Agent and Patient does not tell you the manner or concept encoded by slice, crush, polish.

112. They do not replace syntax

English word order, agreement, prepositions and constructions still have to be learned.

Part XXXI — Semantic Roles Glossary

Semantic role

The semantic relation a participant bears to a predicate/event.

Thematic role

Common synonym for semantic role/thematic relation.

Thematic relation

Relation such as Agent, Patient, Experiencer or Goal between participant and event.

Theta role

Theory-specific term in generative grammar for a predicate-assigned semantic argument role.

Theta-grid

Representation of roles/arguments selected by a predicate.

Argument

Participant selected by a predicate as part of its core structure.

Adjunct

Optional modifier expressing information such as time, manner or location, depending analysis.

Argument structure

Representation of a predicate’s arguments and their semantic/syntactic properties.

Agent

Typically volitional, sentient participant that initiates/controls an action.

Patient

Participant undergoing change or being strongly affected.

Theme

Participant moved, located, possessed or centrally involved without necessarily internal change.

Experiencer

Sentient participant experiencing emotion, perception, cognition or state.

Stimulus

Entity/event that triggers an experience or perception.

Instrument

Entity used to perform an action.

Means

Method/vehicle by which an action is accomplished.

Cause

Non-volitional entity/event causing another event or change.

Recipient

Participant receiving a Theme in transfer.

Beneficiary

Participant for whose benefit an event occurs.

Maleficiary

Participant adversely affected; included in some role systems.

Source

Starting point of motion, transfer or relation.

Goal

Endpoint/destination of motion or transfer.

Path

Route traversed in a motion event.

Location

Place where entity/event is situated.

Possessor

Participant owning/holding an entity.

Possessed

Entity in a possession relation.

Content

Information/thought/message involved in communication or cognition.

Result

Resulting state/entity produced by an event.

Attribute

Property associated with an entity in state/measurement relations.

Value

Magnitude/category assigned to an Attribute.

Actor

General action participant sometimes used when Agent is too strong.

Undergoer

General affected participant sometimes used instead of Patient/Theme distinctions.

Proto-Agent

Dowty’s cluster of Agent-like entailments such as volition, sentience and causation.

Proto-Patient

Dowty’s cluster of Patient-like entailments such as change of state and affectedness.

Volition

Intentional choice/control over an action.

Sentience

Capacity for perception or experience.

Affectedness

Degree to which a participant changes or is impacted by an event.

Causation

Relation in which one participant/event brings about another state/event.

Argument selection

How semantic properties influence which arguments map to grammatical relations.

Linking

Mapping semantic arguments to syntactic positions/functions.

Subject

Grammatical function often but not always associated with Agent-like roles.

Object

Grammatical function often but not always associated with Patient/Theme-like roles.

Oblique

Non-core/marked grammatical relation often expressed by preposition/case.

Voice

Grammatical system altering mapping/perspective of arguments, e.g. active/passive.

Passive

Construction promoting a non-Agent argument, often Patient/Theme, to subject while optionally expressing Agent in a by-phrase.

Causative

Construction/event representation introducing a Causer/Agent causing another event.

Inchoative

Change-of-state use without an overt external Causer, e.g. The glass broke.

Dative alternation

Alternation such as give Arun the book / give the book to Arun.

Locative alternation

Alternation such as load hay onto the wagon / load the wagon with hay.

Resultative

Construction expressing an action plus resulting state, e.g. paint the door red.

Selectional preference

Tendency for a predicate/role to favour particular semantic types of participants.

Semantic case

Meaning-based participant relation historically associated with case grammar.

Case role

Another term used for semantic/thematic role.

Grammatical case

Morphological/syntactic marking such as nominative, accusative or dative.

Role hierarchy

Proposed ordering of semantic roles for prominence/linking; inventories and orders vary by theory.

Frame element

Frame-specific role such as Buyer/Seller, distinct from general semantic roles.

Semantic role labelling

Computational task of identifying predicates, participants and their roles.

Predicate

Expression, often a verb/adjective, that attributes a property/relation/event to arguments.

Valency

Number/types of arguments a lexical predicate selects and how they are realised.

Transitive

Predicate/clause commonly taking subject plus object.

Intransitive

Predicate/clause with no direct object.

Ditransitive

Predicate/construction with three core participants, often Agent, Theme and Recipient.

Unaccusative

Intransitive pattern whose subject is Theme/Patient-like rather than Agent-like under many analyses.

Unergative

Intransitive pattern whose subject is relatively Agent/Actor-like under many analyses.

Experiencer verb

Predicate involving an Experiencer and often a Stimulus/Content.

Motion verb

Predicate involving movement and roles such as Theme, Source, Path and Goal.

Transfer verb

Predicate involving movement/control of a Theme between Giver/Source and Recipient/Goal.

Participant

Entity/concept involved in an event/state/relation.

Event structure

Semantic organisation of events, changes, causes and states underlying predicates.

Part XXXII — Frequently Asked Questions

What is a semantic role?

The relation a participant bears to the event/state expressed by a predicate, such as Agent, Patient or Goal.

What is a thematic role?

Usually another term for semantic role/thematic relation.

What is a theta role?

A theory-specific generative-grammar concept related to the semantic role assigned to an argument by a predicate.

Is Agent the same as subject?

No. Agent is semantic; subject is syntactic. Passive Patient subjects and Experiencer subjects prove the distinction.

Is Patient the same as object?

No. A Patient can become subject in a passive or inchoative clause.

What is an Agent?

A prototypically intentional, sentient participant causing/controlling an action.

What is a Patient?

A participant strongly affected or changed by an event.

What is a Theme?

Usually an entity moved, located or centrally involved without necessarily undergoing internal change.

What is the difference between Patient and Theme?

Patient highlights affectedness/change; Theme often highlights movement/location/neutral participation. Terminology varies.

What is an Experiencer?

A sentient participant who feels, perceives, knows or mentally experiences something.

What is a Stimulus?

The entity/situation causing or triggering an experience.

What is an Instrument?

An entity used by an Agent to carry out an action.

What is a Cause?

A non-volitional entity/event that brings about a result.

What is a Recipient?

Participant that receives a Theme.

What is a Beneficiary?

Participant that benefits from an event.

What is Source?

Starting point of motion/transfer.

What is Goal?

Endpoint/destination of motion/transfer.

What is Path?

Route traversed during motion.

What is Location?

Place where an entity/event is situated.

Can one participant have more than one role property?

Yes. Self-moving humans can be Agent-like and Theme-like; instruments can have Cause-like properties.

Are semantic roles universal?

Recurring semantic relations are useful cross-linguistically, but there is no universally agreed inventory or exact boundaries.

How many semantic roles are there?

No single agreed number. Teaching systems often use a small practical set; theories differ.

Why do linguists disagree on roles?

Some cases are gradient/overlapping, languages package events differently, and theories target different generalisations.

What are Proto-Agent and Proto-Patient?

Dowty’s cluster concepts replacing rigid roles for some argument-selection purposes.

What properties make something Proto-Agent-like?

Volition, sentience, causation, movement and independent existence are typical properties.

What properties make something Proto-Patient-like?

Change of state, affectedness and related event-dependent properties.

What is a theta-grid?

A representation listing the semantic arguments selected by a predicate, e.g. give: Agent, Theme, Recipient.

What is argument structure?

The predicate’s participant structure and its mapping to grammar.

What is valency?

The kinds/number of arguments a lexical item conventionally combines with.

Why does give take three participants?

Its transfer meaning conceptually requires a Giver, Theme and Recipient, though some can be contextually omitted in certain constructions.

Why does put require a location?

Its lexical meaning describes placing a Theme somewhere, so Goal/Location is strongly selected.

Why can break be transitive or intransitive?

The causative form includes a Causer/Agent and Patient; the inchoative form expresses only the changing Patient.

What happens to roles in passive voice?

Roles stay semantic: Patient/Theme may become subject; Agent may move to by-phrase or be omitted.

What is a dative alternation?

A change between double-object and to-phrase structures while core transfer roles remain similar.

What is a locative alternation?

A constructional alternation changing whether Theme or Location is object, often with subtle meaning differences.

How do semantic roles help vocabulary?

They explain verb differences, participant expectations and grammar patterns.

How do roles help reading?

They help answer who did what to whom, who experienced what, and where movement began/ended.

How do roles help writing?

They help select verbs and ensure required participants are present and grammatically organised.

How do roles help context clues?

Knowing expected participants narrows possible meanings and referents.

How do roles help translation?

Roles preserve participant relationships when languages use different word order, cases or constructions.

What is semantic role labelling?

Computational identification of event participants and their roles.

How is semantic role labelling related to FrameNet?

Frame-semantic role labelling can identify frame-specific roles; other systems use broader roles or PropBank-style numbered arguments.

Are FrameNet frame elements semantic roles?

They are semantic participant roles, but generally more frame-specific than Agent/Patient-style abstractions.

Can instruments be subjects?

Yes: The key opened the door. Subject does not imply Agent.

Can Patients be subjects?

Yes: passives and inchoative/unaccusative clauses commonly have Patient/Theme subjects.

Can Experiencers be objects?

Yes: The noise frightened Maya has Maya as Experiencer object.

Can inanimate subjects be Causes?

Yes: storms, explosions, medicines, policies or events can cause changes without volition.

Is the by-phrase always Agent?

Often in passives, but by can express other relations in English; construction/context determines role.

Does with always mark Instrument?

No. It can mark accompaniment or other relations.

Does to always mark Goal?

It often marks spatial Goal or Recipient, but English to has other uses.

Why are roles useful if labels are disputed?

Because the distinctions—intentional actor, affected entity, experiencer, instrument, origin, endpoint—clarify meaning even if terminology differs.

What is the best teaching approach?

Use a small role set, real sentences, contrast with subject/object, and acknowledge borderline cases rather than forcing labels.

What is the main educational takeaway?

Separate semantic participant structure from grammatical position: ask what each entity does in the event, not simply where it appears in the sentence.

Part XXXIII — Semantic Roles by School Stage

113. Primary 1–2: who did what?

Start with simple action scenes. Ask who acted and what was affected: The girl kicked the ball. Use everyday labels before technical terms if needed: doer/action target.

114. Primary 3–4: movement and transfer

Add Giver, Receiver, Thing Given, Source and Goal. Use real objects and arrows to show direction.

115. Primary 5–6: Experiencer and Instrument

Introduce perception/emotion verbs, tools and passive voice. Learners can now see that subject is not always Agent.

116. Secondary 1–2: argument structure

Teach Agent/Patient/Theme/Experiencer/Stimulus/Recipient/Beneficiary/Source/Goal systematically and connect them to verb patterns.

117. Secondary 3–4: alternations and technical language

Use passives, dative alternation, locative alternation, causative/inchoative pairs and subject/object mismatches. Apply roles to science, history and academic reading.

118. JC, university and adult

Introduce theta roles, proto-roles, case marking, frame-specific roles, cross-linguistic variation and computational semantic role labelling.

Part XXXIV — A Teaching Sequence

1. Start with an event learners can act out.
2. Identify participants without labels.
3. Ask what each participant contributes to the event.
4. Introduce Agent/Patient or Experiencer/Stimulus labels.
5. Change the sentence while keeping the event constant.
6. Use active/passive pairs to separate role from syntax.
7. Use converse verbs to change lexical perspective.
8. Add Source/Goal/Path to motion.
9. Add Instrument and Beneficiary.
10. Introduce borderline cases and discuss properties.
11. Connect roles to verb grammar.
12. Transfer the analysis into authentic text.

Part XXXV — Student Operating Manual

1. Find the predicate.
2. List the participants.
3. Ask who intentionally causes or controls the event.
4. Ask what changes or is affected.
5. Ask what moves or is located.
6. Ask who experiences or perceives.
7. Ask what triggers the experience.
8. Ask whether a tool or means is used.
9. Ask who receives or benefits.
10. For motion/transfer, mark Source, Path and Goal.
11. Then identify grammatical subject/object separately.
12. If no label fits cleanly, describe the semantic properties instead of forcing one.

Part XXXVI — Teacher Operating Manual

1. Use semantic roles to explain meaning, not as labels to memorise for their own sake.
2. Keep subject/object terminology separate from Agent/Patient terminology.
3. Use active/passive pairs constantly.
4. Contrast voluntary Agent with non-volitional Cause.
5. Contrast Patient affectedness with Theme movement/location.
6. Teach Experiencer verbs because they disrupt the “subject = doer” shortcut.
7. Use transfer and motion diagrams for Recipient/Source/Goal.
8. Show word-specific valence.
9. Acknowledge theoretical variation.
10. Use proto-role properties for difficult cases.
11. Apply roles to authentic comprehension and writing.
12. Assess whether learners can transfer the analysis to new verbs.

Part XXXVII — Parent Guide

At home, semantic roles can be taught through everyday events without technical terminology. Ask: “Who did it? What changed? Who received it? What did they use? Where did it start and where did it end?” Once the child handles those distinctions easily, introduce the linguistic labels if useful.

Use toys

Move a toy car from a box to a table: car = Theme, box = Source, table = Goal. Let the child be Agent if they move it intentionally.

Use giving

Give a pencil to someone: Giver/Agent, Theme, Recipient. Then reverse the sentence using receive.

Use emotions

“The thunder scared me.” Thunder = Stimulus/Cause; I = Experiencer. This makes clear why the object can be the person feeling something.

Part XXXVIII — First Weak Link Diagnostic

1. Can the learner identify the predicate?
2. Can the learner identify all participants?
3. Can the learner separate grammatical subject from semantic Agent?
4. Can the learner distinguish Agent from Cause?
5. Can the learner distinguish Patient from Theme?
6. Can the learner identify Experiencer and Stimulus?
7. Can the learner distinguish Recipient from Beneficiary?
8. Can the learner mark Source, Path and Goal?
9. Can the learner identify Instrument/Means?
10. Can the learner preserve roles across active/passive?
11. Can the learner preserve roles across lexical perspective changes?
12. Can the learner use role knowledge to choose a correct verb pattern?

Repair the first failed step. If a learner cannot find the predicate, role labels will not help. If they can identify participants but equate subject with Agent, active/passive contrasts are the right repair.

Part XXXIX — Reading Comprehension

119. Pronoun resolution

Role expectations can help identify who pronouns refer to, especially when events involve likely Agents, Experiencers or Recipients.

120. Implicit participants

The door was repaired invites an unexpressed repairer/Agent. Readers understand more than the sentence states overtly.

121. Event coherence

Tracking Agents, Patients and Goals across sentences helps readers build a stable situation model.

122. Technical passages

Scientific prose frequently packs Cause, Instrument, Patient/Theme and Result into dense clauses. Role analysis can unpack them.

123. Ambiguity

A word or noun phrase may support different role assignments, producing different interpretations. Syntax and world knowledge help resolve them.

Part XL — Writing

124. Missing arguments

A writer may use a verb without supplying a participant that English expects. Knowing valency prevents incomplete structures.

125. Wrong perspective verb

Learners often confuse borrow/lend, teach/learn or bring/take because they have the scene but choose the wrong participant perspective.

126. Passive choice

Passive can foreground Patient/Theme when the affected entity matters more than the Agent: The samples were analysed…

127. Agent omission

Academic writing may omit Agents when procedure or result is more important. Role analysis helps readers understand what is implicit.

128. Precision

Different verbs select different roles. Provide X to Y, supply Y with X and give Y X encode similar transfer content through different valence patterns.

Part XLI — Cross-Linguistic Variation

129. Word order varies

Languages can place Agents, Patients and other roles in different orders. Role relations should not be inferred from English order alone.

130. Case marking varies

Some languages mark roles explicitly with case endings; others rely more on word order, agreement or context.

131. Ergative alignment

Some languages group the subject of an intransitive clause with the transitive object in ways unlike English nominative–accusative alignment. This is another reason not to equate “subject” with Agent universally.

132. Experiencer marking varies

Experiencers can be nominative, dative or other grammatical forms across languages. Recent linguistic work continues to examine how experiencer roles map to syntax.

133. Transfer constructions vary

Languages differ in whether Recipients are marked like Goals, objects or by special cases/constructions.

134. Translation rule

Preserve semantic roles first—who acts, changes, experiences, receives, moves—then rebuild the target-language grammar according to its own conventions.

Part XLII — Semantic Roles in Technical Domains

Physics

Useful roles: Cause, Theme, Source, Goal, Instrument. Application: Force causes acceleration; object is affected/moving Theme; instruments measure quantities.

Chemistry

Useful roles: Agent/Cause, Patient/Reactant, Result/Product. Application: Reactions often involve non-volitional Causes rather than human Agents.

Biology

Useful roles: Agent-like Organism, Patient/Theme, Stimulus, Experiencer. Application: Behavioral and physiological descriptions use causal/experiential roles.

Medicine

Useful roles: Experiencer/Patient, Stimulus/Symptom, Agent/Clinician, Instrument. Application: Institutional patient role is not identical to thematic Patient.

Mathematics

Useful roles: Theme/Entity, Value, Source/Goal metaphorically. Application: Roles help only selectively; formal relations may be better handled by frame-specific labels.

Law

Useful roles: Agent/Offender, Patient/Victim, Recipient, Experiencer. Application: Institutional roles refine general roles.

Economics

Useful roles: Agent, Recipient, Theme, Source/Goal. Application: Transfer and exchange verbs map well to semantic roles.

Computing

Useful roles: Agent/User/System, Theme/Data, Source, Goal. Application: Digital transfer language extends motion/transfer roles.

History

Useful roles: Agent/Cause, Patient/Affected Group, Source/Goal. Application: Role analysis clarifies events without reducing complex historical causation to one doer.

Literature

Useful roles: Agent, Experiencer, Stimulus, Theme. Application: Narrative viewpoint often manipulates which participant is foregrounded.

Part XLIII — Advanced Participant-Role Casebook: 70 Verbs

1. admire

Role pattern: Experiencer + Stimulus. Example: Maya admires the painting. Lesson: Subject is Experiencer, not Agent.

2. amuse

Role pattern: Stimulus/Cause + Experiencer. Example: The joke amused Maya. Lesson: Object is Experiencer.

3. annoy

Role pattern: Stimulus/Cause + Experiencer. Example: The noise annoyed Maya. Lesson: Inanimate subject can be Stimulus.

4. appreciate

Role pattern: Experiencer + Stimulus/Theme. Example: Maya appreciates the help. Lesson: Mental-evaluation verb.

5. astonish

Role pattern: Stimulus + Experiencer. Example: The result astonished the team. Lesson: Stimulus subject.

6. attract

Role pattern: Stimulus/Target + Experiencer/Theme. Example: The exhibition attracted visitors. Lesson: Cause-like subject pulls Theme/Experiencer.

7. avoid

Role pattern: Agent/Experiencer + Theme/Stimulus. Example: Maya avoided the crowd. Lesson: Participant intentionally prevents contact.

8. believe

Role pattern: Experiencer + Content. Example: Maya believes the story. Lesson: Subject is Cognizer/Experiencer.

9. belong

Role pattern: Theme/Possessed + Possessor. Example: The book belongs to Maya. Lesson: Subject is possessed entity.

10. benefit

Role pattern: Beneficiary/Experiencer + Source/Theme. Example: Maya benefited from the course. Lesson: Subject receives positive effect, not Agent.

11. blame

Role pattern: Agent/Evaluator + Target + Cause/Offence. Example: Maya blamed Arun for the delay. Lesson: Three semantic participants.

12. borrow

Role pattern: Recipient/Borrower + Theme + Source/Lender. Example: Maya borrowed a book from Arun. Lesson: Subject is Recipient-like despite intentional action.

13. bring

Role pattern: Agent + Theme + Goal/Viewpoint. Example: Maya brought the book to class. Lesson: Deictic viewpoint matters.

14. carry

Role pattern: Agent + Theme + Path/Goal. Example: Maya carried the bag upstairs. Lesson: Theme moves because Agent acts.

15. contain

Role pattern: Container + Theme. Example: The box contains tools. Lesson: Subject is Location/Container-like, not Agent.

16. cost

Role pattern: Goods/Service + Experiencer/Buyer + Money. Example: The ticket cost Maya $50. Lesson: Subject is valued item.

17. cover

Role pattern: Theme/Container-like + Location/Surface. Example: Snow covered the road. Lesson: Subject is material spreading over Patient/Location.

18. cross

Role pattern: Agentive Theme + Path. Example: Maya crossed the bridge. Lesson: Subject both controls and undergoes movement.

19. damage

Role pattern: Agent/Cause + Patient. Example: The storm damaged the roof. Lesson: Cause can replace Agent.

20. deliver

Role pattern: Agent + Theme + Recipient/Goal. Example: The courier delivered the parcel to Maya. Lesson: Transfer/motion roles.

21. depend

Role pattern: Dependent Theme + Source/Condition. Example: Success depends on practice. Lesson: No Agent or Patient.

22. deserve

Role pattern: Experiencer/Recipient-like + Theme/Outcome. Example: Maya deserves praise. Lesson: Stative evaluative relation.

23. die

Role pattern: Patient/Theme. Example: The plant died. Lesson: Subject undergoes terminal change; no external Agent.

24. disappear

Role pattern: Theme. Example: The keys disappeared. Lesson: Subject Theme in change-of-presence event.

25. disgust

Role pattern: Stimulus + Experiencer. Example: The smell disgusted Maya. Lesson: Stimulus subject.

26. enter

Role pattern: Agentive Theme + Goal. Example: Maya entered the room. Lesson: Motion subject.

27. escape

Role pattern: Agentive Theme + Source. Example: Maya escaped the building. Lesson: Subject leaves Source.

28. exist

Role pattern: Theme/Entity + Location. Example: The problem exists in several forms. Lesson: Existential state, no Agent.

29. fall

Role pattern: Theme + Source/Goal/Path. Example: The leaf fell from the tree. Lesson: Non-agentive motion.

30. fit

Role pattern: Theme + Container/Location. Example: The box fits the shelf. Lesson: Relational predicate.

31. follow

Role pattern: Agentive Theme + Path/Reference. Example: Maya followed Arun. Lesson: Subject moves in relation to another Theme.

32. forgive

Role pattern: Experiencer/Evaluator + Target + Offence. Example: Maya forgave Arun for the mistake. Lesson: Social-evaluative roles.

33. frighten

Role pattern: Stimulus/Cause + Experiencer. Example: The dog frightened Maya. Lesson: Object is Experiencer.

34. grow

Role pattern: Theme/Patient. Example: The child grew. Lesson: Change-of-state subject.

35. grow (transitive)

Role pattern: Agent + Theme. Example: The farmer grew tomatoes. Lesson: Same form, different argument structure.

36. hear

Role pattern: Experiencer + Stimulus. Example: Maya heard music. Lesson: Perception without intentional Agentivity.

37. hit

Role pattern: Agent/Cause + Patient/Target + Instrument optional. Example: Maya hit the ball with a bat. Lesson: Patient may move/change rather than internal change.

38. hold

Role pattern: Agent/Possessor + Theme. Example: Maya held the rope. Lesson: Physical control/possession relation.

39. hurt

Role pattern: Cause/Agent + Experiencer/Patient. Example: The fall hurt Maya. Lesson: Affected sentient participant.

40. interest

Role pattern: Stimulus + Experiencer. Example: Science interests Maya. Lesson: Stimulus subject.

41. join

Role pattern: Agentive Theme + Goal/Group. Example: Maya joined the club. Lesson: Membership change expressed like movement to Goal.

42. keep

Role pattern: Agent/Possessor + Theme + State/Location. Example: Maya kept the door open. Lesson: Causative-maintenance relation.

43. know

Role pattern: Experiencer/Knower + Content. Example: Maya knows the answer. Lesson: Stative cognitive relation.

44. lack

Role pattern: Possessor/Entity + Theme. Example: The plan lacks detail. Lesson: Subject is entity missing something, not Agent.

45. leave

Role pattern: Agentive Theme + Source. Example: Maya left the building. Lesson: Departure frame.

46. lend

Role pattern: Agent/Lender + Theme + Recipient. Example: Maya lent Arun a book. Lesson: Converse of borrow.

47. like

Role pattern: Experiencer + Stimulus. Example: Maya likes jazz. Lesson: Experiencer subject.

48. lose

Role pattern: Experiencer/Possessor + Theme. Example: Maya lost her keys. Lesson: Subject affected by possession change; not canonical Agent if accidental.

49. meet

Role pattern: Participant + Co-participant. Example: Maya met Arun. Lesson: Symmetric social relation rather than Agent–Patient.

50. miss

Role pattern: Experiencer + Stimulus/Target. Example: Maya misses Arun. Lesson: Emotion relation.

51. notice

Role pattern: Experiencer + Stimulus. Example: Maya noticed the error. Lesson: Perception/cognition.

52. own

Role pattern: Possessor + Theme. Example: Maya owns the house. Lesson: Possession state.

53. please

Role pattern: Stimulus + Experiencer. Example: The result pleased Maya. Lesson: Experiencer object.

54. pour

Role pattern: Agent + Theme + Goal/Location. Example: Maya poured water into the glass. Lesson: Change of location; container relation.

55. provide

Role pattern: Agent/Source + Theme + Recipient. Example: The school provided students with books. Lesson: Valence differs from give.

56. receive

Role pattern: Recipient + Theme + Source. Example: Maya received a letter from Arun. Lesson: Recipient subject.

57. remain

Role pattern: Theme + State/Location. Example: The door remained open. Lesson: State persistence.

58. remember

Role pattern: Experiencer + Content. Example: Maya remembered the address. Lesson: Mental retrieval.

59. resemble

Role pattern: Theme/Entity + Comparison Entity. Example: Maya resembles her sister. Lesson: Symmetric-like relation; no Agent.

60. return (intransitive)

Role pattern: Agentive Theme + Goal. Example: Maya returned home. Lesson: Motion.

61. return (transitive)

Role pattern: Agent + Theme + Recipient/Goal. Example: Maya returned the book to Arun. Lesson: Caused transfer.

62. rob

Role pattern: Agent/Offender + Victim/Location + Theme optional. Example: The thief robbed Maya of her bag. Lesson: Verb foregrounds victim/source rather than stolen Theme.

63. steal

Role pattern: Agent/Offender + Theme + Source. Example: The thief stole Maya’s bag. Lesson: Verb foregrounds stolen Theme.

64. seem

Role pattern: Theme/Proposition + Experiencer optional. Example: The plan seems risky to Maya. Lesson: Epistemic appearance; no canonical Agent.

65. smell (perception)

Role pattern: Experiencer + Stimulus. Example: Maya smelled the smoke. Lesson: Experiencer subject.

66. smell (property)

Role pattern: Stimulus/Theme + Attribute. Example: The soup smells delicious. Lesson: Subject is stimulus/entity with perceived property.

67. suffer

Role pattern: Experiencer/Patient + Condition. Example: Maya suffered an injury. Lesson: Strong affectedness without Agentivity.

68. surprise

Role pattern: Stimulus + Experiencer. Example: The news surprised Maya. Lesson: Stimulus subject.

69. survive

Role pattern: Experiencer/Theme + Event/Condition. Example: Maya survived the accident. Lesson: Subject persists despite adverse event.

70. taste (perception)

Role pattern: Experiencer + Stimulus. Example: Maya tasted the soup. Lesson: Controlled perception can be more Agent-like.

71. taste (property)

Role pattern: Theme/Stimulus + Attribute. Example: The soup tastes salty. Lesson: Subject carries property; no Agent.

72. teach

Role pattern: Agent/Source + Recipient/Learner + Content. Example: Maya taught Arun algebra. Lesson: Knowledge transfer.

73. understand

Role pattern: Experiencer + Content. Example: Maya understands the problem. Lesson: Cognitive state.

74. use

Role pattern: Agent + Instrument/Theme + Purpose. Example: Maya used a key to open the door. Lesson: Object functions as Instrument but grammatically is object of use.

75. weigh

Role pattern: Theme + Value. Example: The box weighs ten kilograms. Lesson: Measurement relation, no Agent.

76. welcome

Role pattern: Agent/Host + Recipient/Guest. Example: The school welcomed Maya. Lesson: Social reception frame.

77. want

Role pattern: Experiencer/Desirer + Theme/Proposition. Example: Maya wants a break. Lesson: Mental state.

78. watch

Role pattern: Experiencer/Agent-like Observer + Stimulus. Example: Maya watched the match. Lesson: Intentional perception adds Agent-like properties.

79. worry (intransitive)

Role pattern: Experiencer + Content/Stimulus optional. Example: Maya worried about the test. Lesson: Experiencer subject.

80. worry (transitive)

Role pattern: Stimulus + Experiencer. Example: The test worried Maya. Lesson: Stimulus subject.

Part XLIV — Symmetric, Reciprocal and Relational Predicates

135. Meet

Maya met Arun describes a relation in which both participants are co-participants. Labelling one Agent and one Patient misses the semantic symmetry.

136. Marry

Maya married Arun can represent two participants in a reciprocal social relation, though causative/institutional uses create other structures.

137. Resemble

Maya resembles Arun expresses similarity rather than action. Theme/Standard labels may be used, but neither participant is affected.

138. Equal

A equals B is a formal symmetric relation; Agent/Patient terminology is inappropriate.

139. Why this matters

Semantic-role analysis should fit predicate meaning. Not every sentence is an action with a doer and a receiver.

Part XLV — Created, Destroyed and Incrementally Affected Themes

140. Created Theme

Maya wrote a letter. The letter comes into existence through the event. It is a Result/Created Theme rather than a pre-existing Patient.

141. Consumed Theme

Maya ate the sandwich. The sandwich is incrementally consumed and highly Patient-like.

142. Measured Theme

Maya walked three kilometres. The distance phrase measures the event rather than participating as an entity.

143. Incremental Theme

In aspectual semantics, the progress of an event can correspond to how much of a Theme is affected: eating half the sandwich corresponds to partial event completion.

Part XLVI — Role Analysis and Metaphor

Metaphorical language can preserve role structure while shifting domains. The argument attacked the proposal gives an abstract entity Agent/Cause-like behaviour. Readers use conceptual metaphor to understand nonliteral role assignment.

144. Ideas as Agents

The evidence suggests… or The report argues… assigns communication/cognition-like roles to texts.

145. Events as Causes

The crisis forced the government to act. Crisis is a Cause; government is Agent of the subsequent action but affected by causal pressure.

146. Time as mover

Summer is approaching. Abstract time is conceptualised through motion roles.

147. States as Locations

She fell into depression or came out of danger maps abstract states onto Source/Goal/Location structure.

Part XLVII — 80 Semantic Role Classification Drills

1. Maya kicked the ball.

Role analysis: Maya Agent; ball Patient/Theme. Reason: Intentional action affects/moves ball.

2. The ball rolled away.

Role analysis: ball Theme. Reason: Mover without external Agent.

3. The wind rolled the ball downhill.

Role analysis: wind Cause; ball Theme. Reason: Non-volitional causer.

4. Maya watched the bird.

Role analysis: Maya Experiencer/Agent-like Observer; bird Stimulus. Reason: Intentional perception.

5. Maya saw the bird.

Role analysis: Maya Experiencer; bird Stimulus. Reason: Perception without necessarily deliberate attention.

6. The bird attracted Maya’s attention.

Role analysis: bird Stimulus; Maya Experiencer. Reason: Stimulus subject.

7. Maya listened to the music.

Role analysis: Maya Experiencer/Agent-like; music Stimulus. Reason: Intentional auditory attention.

8. Maya heard the music.

Role analysis: Maya Experiencer; music Stimulus. Reason: Less volitional.

9. The music delighted Maya.

Role analysis: music Stimulus; Maya Experiencer. Reason: Emotion causation.

10. Maya enjoyed the music.

Role analysis: Maya Experiencer; music Stimulus/Theme. Reason: Experiencer subject.

11. The medicine cured Maya.

Role analysis: medicine Cause/Instrument-like; Maya Patient/Experiencer. Reason: Inanimate cause produces state change.

12. The doctor cured Maya with the medicine.

Role analysis: doctor Agent; Maya Patient; medicine Instrument/Means. Reason: Agent and treatment means separated.

13. The fire destroyed the house.

Role analysis: fire Cause; house Patient. Reason: Natural/non-volitional cause.

14. Firefighters protected the house.

Role analysis: firefighters Agent; house Patient/Beneficiary-like target. Reason: Protection prevents harm rather than causes change.

15. The wall protected the garden from wind.

Role analysis: wall Instrument/Cause-like Protector; garden Beneficiary/Patient-like; wind Threat/Source. Reason: Inanimate protective subject.

16. Maya unlocked the gate with a code.

Role analysis: Maya Agent; gate Patient; code Instrument. Reason: Intentional change of access state.

17. The code unlocked the gate.

Role analysis: code Instrument/Cause-like; gate Patient. Reason: Instrument as subject.

18. The gate unlocked automatically.

Role analysis: gate Patient/Theme. Reason: Inchoative-like use.

19. Maya passed the note to Arun.

Role analysis: Maya Agent; note Theme; Arun Recipient. Reason: Transfer.

20. Arun got the note.

Role analysis: Arun Recipient; note Theme. Reason: Recipient subject.

21. Maya wrote Arun a note.

Role analysis: Maya Agent; Arun Recipient/Beneficiary; note Created Theme. Reason: Creation plus intended transfer.

22. Maya wrote a note for Arun.

Role analysis: Maya Agent; note Created Theme; Arun Beneficiary. Reason: For-phrase foregrounds benefit.

23. The company offered Maya a job.

Role analysis: company Agent/Source; Maya Recipient; job Theme. Reason: Offer frame.

24. Maya accepted the job.

Role analysis: Maya Agent/Experiencer; job Theme. Reason: Decision relation.

25. The job suited Maya.

Role analysis: job Stimulus/Theme; Maya Experiencer/Standard-like. Reason: Suit relation has no Agent.

26. Maya entered the building.

Role analysis: Maya Agentive Theme; building Goal. Reason: Motion into boundary.

27. Maya exited the building.

Role analysis: Maya Agentive Theme; building Source. Reason: Motion out.

28. The package entered the system.

Role analysis: package Theme; system Goal. Reason: Non-agentive Theme.

29. The signal passed through the filter.

Role analysis: signal Theme; filter Path/Instrument-like environment. Reason: Technical motion metaphor.

30. Water flowed from the tank into the pipe.

Role analysis: water Theme; tank Source; pipe Goal/Path. Reason: Fluid motion.

31. The pipe carried water to the house.

Role analysis: pipe Instrument/Path-like subject; water Theme; house Goal. Reason: Container/means subject.

32. Maya drove the car to school.

Role analysis: Maya Agent; car Theme/Vehicle; school Goal. Reason: Control plus caused motion.

33. The car took Maya to school.

Role analysis: car Instrument/Vehicle/Cause-like; Maya Theme/Passenger; school Goal. Reason: Vehicle subject.

34. The road leads to the village.

Role analysis: road Path/Route subject; village Goal. Reason: Static relational predicate with Path-like subject.

35. The sign points to the exit.

Role analysis: sign Instrument/Indicator; exit Goal/Target. Reason: No intentional Agent.

36. Maya placed the vase on the table.

Role analysis: Maya Agent; vase Theme; table Goal/Location. Reason: Caused placement.

37. The vase stood on the table.

Role analysis: vase Theme; table Location. Reason: Static posture/location.

38. The shelf holds ten books.

Role analysis: shelf Container/Location; books Theme. Reason: Inanimate container subject.

39. The jar contains sugar.

Role analysis: jar Container; sugar Theme. Reason: Containment relation.

40. Sugar filled the jar.

Role analysis: sugar Theme; jar Location/Patient-like container. Reason: Locative alternation changes affectedness.

41. Maya filled the jar with sugar.

Role analysis: Maya Agent; jar Patient/Location; sugar Theme. Reason: Location object construed as affected/completed.

42. Maya sprayed paint on the wall.

Role analysis: Maya Agent; paint Theme; wall Goal/Location. Reason: Theme-object variant.

43. Maya sprayed the wall with paint.

Role analysis: Maya Agent; wall Patient/Location; paint Theme. Reason: Location-object variant.

44. Maya cleared dishes from the table.

Role analysis: Maya Agent; dishes Theme; table Source. Reason: Removal.

45. Maya cleared the table of dishes.

Role analysis: Maya Agent; table Patient/Location; dishes Theme. Reason: Affected location foregrounded.

46. The teacher showed the class a diagram.

Role analysis: teacher Agent; class Recipient/Experiencer; diagram Theme/Stimulus. Reason: Caused perception/information transfer.

47. The class saw the diagram.

Role analysis: class Experiencer; diagram Stimulus. Reason: Perception result.

48. The diagram explains the process.

Role analysis: diagram Stimulus/Information Source; process Content/Theme. Reason: Inanimate information source.

49. The process confused the class.

Role analysis: process Stimulus; class Experiencer. Reason: Cognitive effect.

50. Maya remembered the rule.

Role analysis: Maya Experiencer; rule Content. Reason: Memory state.

51. The rule reminded Maya of the deadline.

Role analysis: rule Stimulus/Cue; Maya Experiencer; deadline Content. Reason: Causative memory access.

52. Maya forgot the deadline.

Role analysis: Maya Experiencer; deadline Content. Reason: Loss/failure of memory access.

53. The smell reminded Maya of home.

Role analysis: smell Stimulus/Cue; Maya Experiencer; home Content. Reason: Association.

54. Maya promised Arun a refund.

Role analysis: Maya Agent/Promiser; Arun Recipient/Addressee; refund Content/Future Theme. Reason: Speech act creates obligation.

55. Arun expected a refund.

Role analysis: Arun Experiencer; refund Content/Theme. Reason: Expectation state.

56. The refund arrived yesterday.

Role analysis: refund Theme; implicit Source/Recipient frame. Reason: Transfer Theme becomes subject.

57. The store refunded Arun.

Role analysis: store Agent/Source; Arun Recipient. Reason: Theme Money implicit.

58. The storm prevented Maya from travelling.

Role analysis: storm Cause/Preventer; Maya Agentive Theme of blocked event. Reason: Causal relation across embedded event.

59. Maya avoided the storm.

Role analysis: Maya Agent; storm Stimulus/Threat/Theme. Reason: Intentional avoidance.

60. The shelter protected Maya from the storm.

Role analysis: shelter Instrument/Cause-like Protector; Maya Beneficiary/Patient; storm Threat. Reason: Multiple causal relations.

61. Maya survived the storm.

Role analysis: Maya Experiencer/Theme; storm Adverse Condition. Reason: Persistence through event.

62. The noise woke Maya.

Role analysis: noise Stimulus/Cause; Maya Patient/Experiencer. Reason: Change of state plus experiencer.

63. Maya woke.

Role analysis: Maya Patient/Experiencer. Reason: Inchoative change.

64. Maya woke Arun.

Role analysis: Maya Agent; Arun Patient/Experiencer. Reason: Causative wake.

65. Maya slept.

Role analysis: Maya Experiencer/Actor-like subject. Reason: One-place state/activity; role label varies.

66. The drug put Maya to sleep.

Role analysis: drug Cause; Maya Patient/Experiencer; sleep Result State. Reason: Causative state change.

67. Maya painted a picture.

Role analysis: Maya Agent; picture Created Theme. Reason: Object created by event.

68. Maya painted the wall.

Role analysis: Maya Agent; wall Patient/Location. Reason: Object already exists and is affected.

69. Maya painted the wall blue.

Role analysis: Maya Agent; wall Patient; blue Result State. Reason: Resultative.

70. The wall turned blue.

Role analysis: wall Patient/Theme; blue Result State. Reason: No overt Agent.

71. Maya built a bridge.

Role analysis: Maya Agent; bridge Created Theme. Reason: Creation.

72. The bridge spans the river.

Role analysis: bridge Theme/Structure; river Path/Location. Reason: Static relation.

73. The river flooded the road.

Role analysis: river/water Cause/Theme; road Patient/Location. Reason: Natural cause and spreading Theme.

74. The road flooded.

Role analysis: road Location/Patient-like subject. Reason: Change-of-state/location construction.

75. The alarm alerted Maya.

Role analysis: alarm Stimulus/Cause; Maya Experiencer/Recipient. Reason: Warning communication.

76. Maya alerted the team.

Role analysis: Maya Agent/Source; team Recipient/Experiencer. Reason: Human Agent warning.

77. The team responded to the alarm.

Role analysis: team Agent/Experiencer; alarm Stimulus. Reason: Response event.

78. Maya benefited from the training.

Role analysis: Maya Beneficiary/Experiencer; training Source/Cause. Reason: Positive effect.

79. The training benefited Maya.

Role analysis: training Cause/Source; Maya Beneficiary. Reason: Perspective alternation.

80. The decision affected Maya.

Role analysis: decision Cause/Stimulus; Maya Patient/Experiencer. Reason: Broad affectedness.

81. Maya influenced the decision.

Role analysis: Maya Agent/Cause; decision Theme/Result target. Reason: Reverse causal direction.

82. The evidence supported the claim.

Role analysis: evidence Source/Cause-like support; claim Theme/Proposition. Reason: Abstract semantic roles in argument.

83. The claim relies on evidence.

Role analysis: claim Theme/Dependent; evidence Source/Basis. Reason: Relational frame, no Agent.

84. The data contradict the hypothesis.

Role analysis: data Evidence/Stimulus; hypothesis Target/Content. Reason: Logical relation, not action Agent–Patient.

85. The hypothesis predicts the result.

Role analysis: hypothesis Information Source/Model; result Content. Reason: Inanimate abstract subject.

86. The result confirmed the hypothesis.

Role analysis: result Evidence/Stimulus; hypothesis Content/Target. Reason: Epistemic relation.

87. The model generated predictions.

Role analysis: model Instrument/System/Cause-like; predictions Created Theme. Reason: Technical computational frame.

88. Researchers trained the model.

Role analysis: researchers Agent; model Patient/Theme. Reason: Technical use of train has model as affected entity.

89. The model learned a pattern.

Role analysis: model Experiencer/Agent-like metaphorical system; pattern Content. Reason: Technical metaphor stretches human role labels.

Part XLVIII — When the Best Answer Is “Borderline”

Role analysis should increase semantic precision, not create false certainty. Several recurring structures legitimately support more than one label depending the theoretical framework.

Self-motion subject

Plausible analysis: Agent + Theme properties. Teaching rule: Maya walks/runs: intentional controller and moving entity.

Instrument subject

Plausible analysis: Instrument + Cause properties. Teaching rule: The key opened the door.

Natural force

Plausible analysis: Cause with some Agent-like causation. Teaching rule: The wind opened the gate.

Experiencer subject

Plausible analysis: Sentient Proto-Agent properties without action causation. Teaching rule: Maya fears spiders.

Created object

Plausible analysis: Theme/Result, sometimes Patient terminology avoided. Teaching rule: Maya wrote a letter.

Consumed object

Plausible analysis: Theme + highly affected Patient properties. Teaching rule: Maya ate the apple.

Container object in locative alternation

Plausible analysis: Location + Patient-like affectedness. Teaching rule: Maya loaded the wagon with hay.

Medical patient

Plausible analysis: Institutional Patient may be Experiencer/Recipient thematically. Teaching rule: The patient received treatment.

System/model as subject

Plausible analysis: Instrument/Cause/Agent metaphor. Teaching rule: The algorithm classified the image.

Abstract evidence subject

Plausible analysis: Stimulus/Cause/Information Source. Teaching rule: The evidence convinced the jury.