This chapter identifies aspects of thought structure that are shared by all languages regardless of whatever structure may be imposed by the semantic resources of a particular language. We can begin by recognizing a fundamental dichotomy between the content of a thought and whatever affect (emotion, mood, attitude) may accompany that content. Presumably content is processed above all in the neocortex while affect is processed in subcortical areas. Both aspects of thought are represented constantly in language, content above all in “segmental” sounds and affect above all in prosody (see Chapter 8). This chapter is concerned with content, where we can recognize a distinction between ideas, the basic building blocks of thought, and the way those ideas are oriented. The distinction between ideas and their orientations will be fundamental to everything that follows in this book.
What all languages share suggests that ideas can be assigned to three basic types that are presumably derived from three varieties of human experience. These three idea types are either “wired into” every human brain or reflect universal human experiences, but those two sources need not be in conflict. We all have similar bodies and similar basic needs, we all inhabit the same planet, and we all share fundamental aspects of our lives like eating and sleeping. Innate wiring and common experiences must both have been at work as our brains have adapted genetically and practically to universal features of the human condition.
Language universals, then, suggest that we experience ideas of events (things that happen) and ideas of states (the way things are). An event is a dynamic experience involving some kind of change. It may be an action someone performs (he ran away) or a change in the state of someone or something (he got sick, it disappeared). A state is the way something is, a more stable condition (he’s sick, it’s missing). Ideas of events and ideas of states are almost always accompanied by a third kind of idea that includes people, objects, and abstractions that initiate, are affected by, or otherwise participate in them. Ideas of this third type are sometimes called “referents” or simply “things.” Here they will be called ideas of “entities.” In the above examples ideas of entities were expressed by the pronouns he and it.
Because events and states share certain properties that contrast them with entities, it would be useful to have a term that combined the two, but unfortunately the English language does not offer such a term. “State of affairs” or “situation” are hardly appropriate terms for an event, and perhaps we can do no better than to settle on simply “event or state” or more briefly “event/state.” It is awkward also to refer constantly to event ideas, state ideas, and entity ideas, so “events,” “states,” and “entities” will be used with the understanding that they refer, not to events, states, and entities themselves, but to types of ideas.
Events, states, and entities can be imagined as situated within a multidimensional matrix of orientations in space, time, epistemology, emotions, social interactions, and the ever-changing context. Every language has its own ways of verbalizing ideas and their orientations: ideas from an inventory of verbs, nouns, and perhaps adjectives, orientations with inflections of verbs and nouns or with separate particles. Language-specific ways of verbalizing ideas and their orientations constitute the unique semantic resources of each language.
Distinguishing the verbs that express events from the nouns that express entities appears to be universal in the languages of the world. Its universality has occasionally been questioned, for example with respect to languages spoken in the northwestern part of North America belonging to the Wakashan and Salishan language families. William Jacobsen (Reference Jacobsen and Efrat1979) surveyed this issue for Wakashan, Van Eijk and Hess (Reference Van Eijk and Hess1986) for Salishan, and both studies concluded that the languages of the two families do make such a distinction. From a thought-based perspective it is difficult to imagine that speakers of a Wakashan or Salishan language would be incapable of distinguishing an event from an entity.
It is true that in many languages the same sound is sometimes used for both. Examples in English are words like “cut,” “water,” “hope,” and many others. English speakers know which they are thinking of, and probably that is true of the speakers of all languages. There are a few experiences, like lightning, whose status may be uncertain, but they only highlight the near universality of the distinction.
Several linguists have noticed that it is impossible to experience the idea of a particular event or state without including the idea of a particular entity.Footnote 3 In contrast, it is possible to experience the idea of a particular entity without assigning it to a particular event or state. As noted by Givón, “a verb-coded event (‘break’, ‘walk’, ‘talk to’, ‘give’, etc.) cannot be experienced – makes no sense – independently of its noun-coded participants” (Givón Reference Givón2001: 53). Ronald Langacker also noticed that “an event is conceptually dependent; it cannot be conceptualized without conceptualizing the participants who interact to constitute it,” whereas “an object is conceptually autonomous, in the sense that we can conceptualize it independently of its participation in any event” (Langacker Reference Langacker2008: 1004). William Croft made the same observation: “For example, one cannot conceive of an action such as running without the involvement of a runner, or of a property such as height without something that is tall. On the other hand, one can conceive of a chair or a dog without the involvement of another concept” (Croft Reference Croft2001: 87).
As discussed further in Chapter 8, as thoughts flow through consciousness the focus of consciousness changes rapidly, its transient quality reflected in an obvious way in language, where the idea of an event or state is typically expressed for only a second or two before the speaker moves on. We can observe this movement in the following example, which includes event and state designations and time measurements.
(1)
For one thing I had uh, state 1.0 (2)
such a tiny apartment. state 0.9 (3)
Um, filler 0.5 (4)
when we … went back to New York, event 1.9 (5)
um .. I always kept a studio there. state 1.6 (6)
And .. of course that was small, state 1.6 (7)
and then, link 0.6 (8)
we got another studio.. in the, event 1.7 (9)
in .. our neighborhood that we líked, state 1.6 (10)
in the Village, state 0.6 (11)
and we moved into that, event 1.0 (12)
and we just got so busy, event 1.1 (13)
we didn’t get around to- event 1.0 (14)
moving into a real apartment, event 1.3 (15)
until this past year. state 1.2
Unlike events, entities may persist for a longer time in both thought and language. In this example the idea of the speaker herself, either alone or including her husband, was expressed as I or we and was present in eight of the fifteen phrases. The idea of the first apartment was present in three (a tiny apartment, a studio, and that), while the idea of New York was present in two (New York and there), as was the idea of the second apartment (another studio and that) as well as the idea of Greenwich Village (our neighborhood and the Village). The event expressed with the words went back in line 4 is inconceivable without the inclusion of the speaker, her husband, and New York. But she could certainly think of herself, her husband, or New York as participants in any number of events. Events, then, differ from entities in two ways: the transience in thought of events compared with the frequent persistence of entities and the fact that events necessarily include entities but not vice versa.
States have properties of their own. They share with events the necessary inclusion of entities, but their time span is not, as it is with events, independent of the time span of those entities. The time span of a state typically mirrors that of the entity with which it is associated. In this example the state of being small must have characterized the New York apartment whenever it was thought of or mentioned. The brief time span of an event is independent of the persistence of an included entity, but that is not true of the time span of a state.
How Entities Relate to Events and States
Having seen how thoughts include ideas of events and states and the entities that participate in them, it remains to consider just how entities may participate in those events and states, or what have been called their “participant roles” (cf. Mithun and Chafe Reference Mithun and Chafe1999). We can focus here on “core” roles, as opposed to “oblique” roles that are less central to an event or state. Core roles are required by a particular event or state whereas oblique roles are optional. The core roles in English are those expressed syntactically as subjects and objects, while oblique roles are typically expressed with prepositional phrases. Also important here is the distinction between intransitive clauses with only one core role and transitive clauses with two.
Subjects and Objects
In languages like English the single core participant of an intransitive clause plays the role of subject and the two core roles of a transitive clause are subject and object. These names refer to syntactic roles and we can ask whether they have a semantic basis. A subject is sometimes said to function semantically as a “topic” or “what a clause is about.” An early discussion of such proposals can be found in Jespersen (Reference Jespersen1924: 146). Lambrecht (Reference Lambrecht1994) provides a more recent discussion of topicality, aboutness, and subjecthood. In Chafe (Reference Chafe1994: 82–92) the metaphor of a “starting point” captured the observation that a subject establishes a point of departure for whatever is included in the rest of a clause.
Subjects exhibit properties that might be expected of starting points. Most subjects, for example, are given (judged to be already active in a listener’s consciousness), a smaller proportion are at least accessible (judged to be semi-active for a listener), and only a few are new (judged to be inactive for a listener), and then only under limited circumstances. In a sample of conversational language discussed in Chafe (Reference Chafe1994: 85–91), 81 percent of the subjects were given, 16 percent were accessible, and only three percent were new. An entity that is already in active consciousness makes the best point of departure for adding something more. Subjects are also usually identifiable (or “definite”). When there is a choice, people typically select as subjects first persons over second or third, humans over non-humans, and animates over inanimates. People tend to start from their own point of view or from the point of view of someone with whom they can easily empathize. Agents (see below) are also more natural starting points than patients. No one of these properties alone determines what will be a subject, but each contributes to that role.
The following excerpt from a family conversation illustrates subject use. The speaker was talking about a windy day of skiing at an area in the Sierra Nevada mountains of California called Boreal Ridge. The subjects are italicized.
(1) Doug and I went skìing at Boreal,
(2) It was windy that day,
(3) when we get to the top of the chairlift,
(4) and it was like you were a clapper on a bell you know,
(5) when you go to get off,
(6) a huge gust came by,
(7) just as we were landing,
(8) we were gonna like,
(9) you know,
(10) we were coming ìn like this,
(11) and you know,
(12) so we’d wait to,
(13) All right now!
(14) We jumped off.
(15) It was great.
The starting point throughout most of this excerpt was the idea of the speaker and his brother Doug, introduced first with the full noun phrase Doug and I and then continued with the pronoun we. This idea was accessible in line 1 on the basis of what had been said earlier in the conversation; in the rest of the excerpt it was given. Other starting points made brief appearances. The word it appears as the subject of being windy in 2 and of was in 4. Generalized you appeared in 4 and 5, and in 15 there was an it that summarized the whole experience. Another you appeared as the subject of the formulaic you know in 9 and 11. The subject of 6 is of special interest because a huge gust expressed new information that was introduced with the “presentative” predicate came by (Chafe Reference Chafe1994: 111–113). A new subject like this usually expresses subsidiary information that fails to be activated again (Chafe Reference Chafe1994: 90–91).
Subjects may be expressed with “nominative” case marking on nouns and pronouns or else with no case marking at all. In languages that have a more or less fixed word order the orders in which the subject precedes the object (SOV, SVO, VSO) predominate over orders in which the subject follows the object, another property that might be expected of a starting point.
The semantic role of object is mentioned below.
Ergatives and Absolutives
A different way of assigning participant roles can be illustrated with examples from Central Alaskan Yup’ik, a language of the Eskimo-Aleut family that is spoken in southwestern Alaska (Jacobson Reference Jacobson1995). In Yup’ik both the marking of case on nouns and the pronominal suffixes on verbs show the kind of patterning that has been termed ergative–absolutive. A survey of views regarding this pattern is available in Frans Plank (Reference Plank and Plank1979). The following examples depart from colloquial usage but they serve the purpose of illustrating the ergative–absolutive pattern.
With an intransitive clause the single argument is in the absolutive case:
(1)
Arnaq ikayuutuq. arnaq ikayuu-tu-q woman (absolutive) help-INTRANSITIVE.INDICATIVE-3rd.SINGULAR “The woman helped.”
With a transitive clause, what would be a subject in English is in the ergative case and the other core argument is absolutive:
(2)
Angutem arnaq ikayuraa. angutem arnaq ikayur-a-aa man (ergative) woman (absolutive) help-TRANSITIVE.INDICATIVE-3rd.SINGULAR/3rd.SINGULAR “The man helped the woman.”
What are the semantic roles that underlie absolutive and ergative participants? An answer suggested by several linguists, including Aleksandr Kibrik, Edward Keenan, and Marianne Mithun, is illustrated by these Yup’ik examples. The ergative–absolutive pattern is dominated by the absolutive, the role that is usually unmarked formally and is obligatorily present in all clauses. Kibrik (Reference Kibrik and Plank1979: 66) characterized the absolutive as the “closest participant in the situation … an actant [i.e. core participant] who directly takes part in it.” Keenan (Reference Keenan1984: 200–205) characterized it in terms of “bondedness to the verb.” Mithun (Reference Mithun, Fox and Hopper1994: 255) called it the “participant most immediately or directly involved in the event or state.”
Why should the woman in 1 be assigned the same role as the woman in 2? What they have in common is suggested by a pair of English sentences:
(3) The glass broke.
(4) Henry broke the glass.
In both 3 and 4 the glass can be seen as the most involved participant in the breaking event as opposed to Henry in 4. If this were an ergative–absolutive language the glass would play the absolutive role and Henry the ergative. Keenan (Reference Keenan1984: 201) noted that the meanings of verbs often depend on the nature of the absolutive nouns that are associated with them. For example, the meaning assigned to the intransitive verb run varies according to who or what is running (the absolutive participant) in the following examples:
(1) John is still running.
(2) My watch is still running.
(3) My nose is still running.
(4) The Braque exhibition is still running.
This same dependence on the meaning of the absolutive can be seen with transitive verbs as well. The meaning of cut varies according to what is cut (again the absolutive participant) in:
(1) He cut his foot.
(2) He cut the lawn.
(3) He cut the paper.
(4) He cut his whisky with water.
The semantic role of an ergative participant (“he” in these last examples) is mentioned below.
Agents and Patients
A third way of assigning participant roles has received less attention and is often misunderstood. The Seneca language, where this arrangement predominates, provides an illustration. The sentence below, taken from a Seneca folktale, illustrates both agent and patient roles. In the English translation the words I and she are both subjects, but in the Seneca the prefix wag- (from wak-) in sawagadöswe’dë’ “I get hungry again” expresses a first person patient of becoming hungry, whereas the prefix ag- (from yak-) in wa’a:gë’ “she said” expresses a feminine agent of the act of saying.Footnote 4
| Da’áö́h | sawagadöswe’dë’ | wë:doh | gyö’öh | wa’a:gë’. |
| ta’áö́h | s-a-wak-atöswe’t-ë’ | wëtöh | kyö’öh | wa’-yak-ë-’ |
| impossible | REP-FAC-1.PAT-get.hungry-PUN | ever | hearsay | FAC-FEM.AGT-say-PUN |
| “I can’t ever get hungry again she said.” | ||||
Details of this agent–patient pattern can vary with the language, but in every case there is a characterization of entities according to their direct roles in events and states. Entities that are treated as agents willfully instigate events over which they have control. Those treated as patients are not in control but suffer the effects of events that befall them or of states that apply to them. A discussion of the features underlying systems of this type and the ways they can vary across languages is available in Mithun (Reference Mithun1991). These roles are determined by the semantics of the event or state itself, and not by discourse factors as is the case with starting points and absolutives. It is not surprising, then, that the agent–patient pattern appears most often as a part of verb morphology, typically in the pronominal affixes on verbs as in the above example.
Two questions remain: what is the semantic role of the object in the subject–object arrangement, and what is the semantic role of the ergative in the ergative–absolutive arrangement? The check marks in Table 5.1 suggest answers, namely that objects play the role of the most involved participant, and that ergatives play the role of agent. These parallels capture close resemblances across the three systems. One might also ask whether these participant roles are universal aspects of thought. The fact that they vary from language to language suggests that they are not intrinsic to thought but are assigned differently by the differing semantics of different languages. The point to be stressed is that subjects, objects, ergatives, absolutives, agents, and patients all have a semantic basis and are not merely arbitrary syntactic roles.
Table 5.1 The syntax and semantics of participant roles
 |
Summary
On the basis of linguistic universals we can distinguish ideas (the building blocks of thought) from orientations of those ideas. Ideas may be ideas of events, states, or entities. Events and states demand the inclusion of entities whereas the reverse is not true; an entity may stand alone. In the flow of thought, events are usually transient whereas entities often enjoy longer time spans. States may differ from events by adhering to the time span of the entities with which they are associated. We then looked at ways in which entities are related to events and states, or participant roles, and contrasted the properties of subject–object, ergative–absolutive, and agent–patient systems.
