linguistic perspective (i.e., on the basis of purely linguistic indications), however, a temporal link can be conceived of as being merely epiphenomenal. In the first conjunct of example (21), the past-tense form vertrok ‘left, was leaving’ co-occurs with both the past time adverb gisteravond ‘last night’ and the future time adverb morgen ‘tomorrow’. This relationship contrasts with the relationship between the time adverb and tense form in the second conjunct, which we turn to now. In the second conjunct of (21) and in (22), the situations referred to are situated in the future. (22) Tomorrow is Sunday. In the second conjunct of (21), the present-tense form vertrek ‘leave, are leaving’ co-occurs with both the present time adverb nu ‘now’ and the future time adverb overmorgen ‘the day after tomorrow’. In English, present-tense forms can be used for future situations if ‘‘the constitution, order, schedule, habit of things is such that the occurrence can be expected to take place’’ (Calver 1946: 323; see also Langacker 1991: 263–66; 2001b), as is the case in (22). In Dutch, the present-tense form has a somewhat wider range of uses: it signals that the situation in question is (considered to be) on the agenda. Let us now consider the non-time-based tense analyses of Langacker (1978, 1991), Brisard (1999), and Janssen (1989, 1991, 1993, 2002). Langacker (1991: 242–46) holds that present and past tense are grounding predications which relate a situation to an ‘‘epistemic’’ domain. Since the situation is regarded as directly accessible to the language users, they accept it as old informa- tion. Langacker relates this acceptance of a directly accessible situation as old infor- mation to the definiteness (mental contact with a referent) of demonstratives. In particular, he relates the proximal/distal contrast in demonstratives to the present/ past distinction. In other words, in his view, the English present- and past-tense forms signal a proximal/distal distinction, whereby the present/pasttime interpre- tation is merely one possible instantiation of this distinction, albeit the prototypical one. The proximal form indicates that the situation described is actual and imme- diate to the ground (in particular the time of speaking). The distal form indicates the nonimmediacy ofthe situation described: ‘‘Distance within reality normally amounts to ‘past tense’ ’’ (Langacker 1994: 141). The term ‘‘distal’’ suggests distance from the ground ‘‘not in a temporal but in an epistemic sense’’ (Langacker 1978: 869). Brisard (1999, 2002), elaborating on Langacker’s idea of the epistemic status of tenses, defines the present tense in terms of givenness and presence (immediacy) (1999: 367) and the past tense in terms of givenness and nonpresence (absence/ nonimmediacy) (1999: 353). He explicitly rejects Langacker’s distinctive terms ‘‘proximal’’ and ‘‘distal’’ in order to avoid ‘‘a metaphorical interpretation of ‘dis- tance’, in which temporal distance (‘pastness’) is shifted to the domain of epistemic reasoning (‘hypotheticalness’)’’ (Brisard 1999: 235). The notion of givenness is the function that is assigned to the notion ‘‘time of reference’’ in other analyses. Janssen (1987, 2002) relates tense to other types of deixis, such as demonstra- tives. 12 He examines how a situation to be described is contextualized by means of a deictic element, whereby he assumes a vantage point from which the language 810 ronny boogaart and theo janssen user, and possibly the addressee, surveys a mental field of vision. In the case of tense, the mental field of vision is divided into two contexts-of-situation, such that a verb in the present-tense form signals ‘verb-in-this-context-of-situation’, whereas a verb in the past-tense form signals ‘verb-in-that-context-of-situation’. The Reichenbachian relation between s and r is, in Janssen’s analysis, ac- counted for by the relation of the speaker’s, and possibly the addressee’s, vantage point (usually the speech situation) with either this-context-of-situation (the current mental field’s region of focal concern) or that-context-of-situation (the current mental field’s region of disfocal concern). The situation described by means of a tensed clause is assumed to occupy the relevant context-of-situation. Thus, the relevant context-of-situation is considered the situation’s direct frame of reference. Since Janssen replaces r by a situational frame of reference, his approach to the three situations of sentence (2), which are temporally different but situationally related, allows for a meaningful coherent connection between the situations in- volved. This connection follows from the situated interpretation by the addressee. The three situations in sentence (2) can be conceived of as related to one single situational frame of reference, namely the one indicated by means of the relative temporal adverb when (Janssen 1998). The postulation of the reference point, or frame, in the semantics of tense is not undisputed. The notion serves primarily to explain why situations presented by means of a finite, or tensed, form are conceived of as definite: there has to be a unique relationship with the time or frame of reference given by the context or situation. In this view, the notion can be used to distinguish, for instance, between the interpretation of the simple past and the present perfect in languages like English and Dutch. In the present perfect construction, the main verb is a nonfinite form (past participle); therefore, the construction can be used to present situations that are not linked to an already given time or frame of reference, but that are ‘all new’. However, some tensed forms, such as the simple past tense in English, may also signal ‘indefinite past’. See (23) and (24). (23) Cicero was executed by Marcus Antonius. (Michaelis 1998: 225) (24) What happened to your sister? She bought a gold-mine. (adapted from Heny 1982: 134) In order for (23) and (24) to be felicitous, no time or frame has to be given beforehand. For such sentences, therefore, a one-dimensional analysis of the past tense (whereby the situation precedes the evaluative situation) seems to be suffi- cient. Given such cases, in which no r seems to be needed, it is questionable whether ‘‘definiteness’’ and the notion of r can be considered to be present in the semantics of tense or whether, as Michaelis (1998: 226) argues, the English simple past tense is unmarked with respect to this feature. Boogaart (1999) has argued that the presence or absence of r, as well as the ‘‘definiteness’’ of tense, is a matter of aspect rather than tense. The imperfective past requires a previously given, or inferable, reference time, whereas the perfective past—which, in Boogaart’s view, includes eventive clauses in English such as (23) and (24)—is compatible with such a definite reading, but does not require it. tense and aspect 811 3. Aspect Whereas tense locates a situation with respect to the evaluative situation (usually the time of speech), aspect does not serve any such deictic, or grounding, function. Rather than linking the situation externally to the discourse’s ground, aspect con- cerns the internal temporal structure of situations (Comrie 1976). More specifically, aspect indicates whether a situation is conceptualized as unbounded (imperfective aspect) or as bounded (perfective aspect). 13 Before addressing the form (section 3.2) and meaning (section 3.3) of aspect, we will first make some remarks on the interaction between aspect and tense (section 3.1). 3.1. Aspect and Tense In order to determine whether a situation is either ‘‘in progress’’ or ‘‘completed’’ (aspect), some evaluative situation is arguably needed, just as it is the case for the category of tense. However, the evaluative situation needed for the interpretation of aspect does not necessarily consist of the ground of the discourse (the point of speech). For instance, a situation may be ongoing at a point in time preceding the point of speech (imperfective past), at the point of speech itself (imperfective present), or at a point in time which is future with respect to the point of speech (imperfective future): I was reading a book, I am reading a book, I will be reading a book. Therefore, aspect is not considered a deictic category and is, in principle, independent of tense. In terms of Cognitive Grammar, both tenseand aspect delimit what counts as the profiled part of a situation, but they do so at different levels. For instance, the English progressive (aspect) imposes an ‘‘immediate scope’’ that ex- cludes the end point of a situation; tense marking, as either present or past, imposes its own ‘‘immediate scope’’ which is located with respect to the time of speaking (Langacker 2001b: 259–61). In spite of this, the categories of aspect and tense do interact. This is most clearly illustrated by the incompatibility of present tense and perfective aspect: a situation cannot be simultaneously complete (perfective aspect) and valid at the moment of speech (present tense). In English, events (as opposed to states) pre- sented by means of a simple tense get a perfective reading. The event is interpreted as a completed whole. If the tense is past, as in (25a), the tense-aspect combination does not constitute a problem. But if the tense is present, as in (25b), the sense of completion (aspect) clashes with the semantic contribution of the present tense. Therefore, (25b) is infelicitous, unless it receives a habitual reading. 14 (25) a. He learnt the poem. b. He learns the poem. (Langacker 1999: 223) Similarly, perfective forms are usually interpreted as referring to the past if tense marking is lacking (see Moore 2000). 15 In their large-scale typological research, 812 ronny boogaart and theo janssen Dahl (1985) and Bybee, Perkins, and Pagliuca (1994) likewise found that perfective forms typically describe events that are in the past. Thus, even though aspect is not a deictic category, aspectual information sometimes enables one to infer temporal location with respect to the ground. 3.2. Form Information about the internal temporal structure of situations, that is, aspectual information, can be expressed in different ways. The traditional literature on aspect often makes a distinction between aspectual information present in lexical items, most notably verbs, and aspectual information as expressed by language-specific grammatical means, such as the perfective-imperfective morphology in Slavic languages. The first type of aspectual information is referred to as lexical aspect or Aktionsart. The second type is labeled as grammatical aspect or simply aspect (see Binnick 1991: 135–214 on the history of the terms). The distinction between lexical aspect and grammatical aspect raises interesting questions: Is it necessary to make the distinction at all? If one does make the distinction, then how do the two sub- systems of aspect interact? And where do constructions like the English progressive fit in the picture? Vendler (1967) made an influential contribution to the study of lexical aspect when he proposed his ‘‘four time schemata implied by the use of English verbs’’ (144). He divided the verbs of English into four classes depending on (i) whether or not the situation as expressed by the verb has duration, (ii) whether the situation involves change, and (iii) whether the situation is telic, that is, has an ‘‘inherent end point.’’ This results in the four so-called Vendler classes, given in (26) with some of Vendler’s own examples. (26) a. STATE (durative, no change, atelic), e.g., have, possess, know, love, hate b. ACTIVITY (durative, change, atelic), e.g., walk, swim, push, pull c. ACCOMPLISHMENT (durative, change, telic), e.g., paint a picture, build a house d. ACHIEVEMENT (nondurative, change, telic), e.g., recognize, stop, start The class to which a given verb belongs can be detected by ‘‘diagnostic tests,’’ such as its compatibility with the progressive—which distinguishes states from nonstates—or with adverbial phrases as in/for x hours (see Dowty 1979). The latter is illustrated in (27) and (28). (27) a. He walked for two hours. b. ?He walked a mile for two hours. (28) a. ?He walked in two hours. b. He walked a mile in two hours. Atelic predicates like walk in (27a) can be combined with for-adverbials, whereas telic predicates such as walk a mile in (27b) cannot, unless they receive an iterative reading. The reverse is true for adverbials of the in x hours kind: they are fine with tense and aspect 813 telic predicates, as in (28b), but marked with telic predicates, as in (28a), which can only get a kind of inceptive reading. Even though Vendler’s classes are referred to as verb classes, it should be ap- parent that in order to determine Aktionsart, looking at verbs alone does not suffice. According to the test in (27)–(28), the verb walk by itself is atelic, whereas the predicate walk a mile is telic. But if other elements in the clause co-determine ‘‘lexical’’ aspect, this raises questions as to the usefulness of classifying verbs as such. The problem is all the more urgent since the influence of other elements in the clause is not restricted to the direct object. The nature of the subject should equally be taken into account (e.g., Lotsa people walked a mile for two hours). This suggests that Aktionsart is a property of complete clauses rather than a property of verbs or predicates. Verkuyl (1993) has shown that it can be built up compositionally out of the temporal information given by the verb and the nontemporal information provided by its arguments, in other words, whether or not they designate a ‘‘spec- ified quantity.’’ Independently of the Vendler-Dowty tradition, which dealt with lexical aspect, there was a long-standing tradition of studies into the perfective/imperfective distinction as expressed in, for instance, Slavic languages and in the past tenses of the Romance languages. Early on, Garey (1957) made clear that, semantically, the grammatical distinction between perfective and imperfective aspect cannot be equated with the telic/atelic distinction assumed by the lexical tradition. More specifically, a telic situation has an inherent end point, independently of whether it is presented by means of a perfective past, like the French passe ´ simple and passe ´ compose ´ , or by means of an imperfective past, such as the French imparfait. The independence of Aktionsart and aspect is shown in (29) (Garey 1957: 106). 16 (29) Imperfective Perfective Telic Pierre arrivait Pierre est arrive ´ Atelic Pierre jouait Pierre a joue ´ The situation described by the French verb arriver ‘to arrive’ is telic, but it may be presented by either the imperfective past imparfait (arrivait) or the perfective past passe ´ compose ´ (est arrive ´ ). Whereas a perfective past sets up a telic situation as a complete whole, an imperfective past presents a situation from an internal view- point, as being in progress at a particular point in time, and it does not necessarily imply that the inherent end point of the telic situation was actually reached. (However, in the case of nondurative situation types, such as expressed by arriver, the begin point and the end point of the situation more or less coincide so it is hard to imagine a context for Pierre arrivait in which Pierre did not actually arrive.) Likewise, the verb jouer ‘to play’ presents an atelic situation, but such a situation may still be presented either imperfectively ( jouait) or perfectively (a joue ´ ). Atelic situation types do not, at the level of Aktionsart, have a natural end point, but when such a state of affairs is presented by means of a perfective form, it is understood to have ended at some, relatively arbitrary, point in time. The distinction between telicity and perfectivity, and thus between lexical aspect and grammatical aspect, is 814 ronny boogaart and theo janssen well known in the aspect literature. Morphological expression by means of affixes, such as in Slavic languages, is not considered a precondition for talking about (grammatical) aspect: the term is used also for periphrastic constructions like the French passe ´ compose ´ in (29) and the English progressive. The latter construction is usually regarded as a subcategory, or a restricted application, of the category imperfective aspect (Comrie 1976), the main restriction being that it cannot be used with most stative predicates (?He is having blond hair). In the cognitive linguistic literature, the distinction between lexical and gram- matical aspect is usually not made at all, and traditional labels are used to denote different concepts. We will briefly discuss the approaches to aspect advocated by Langacker, Croft, and Michaelis. According to Langacker (1999: 223), ‘‘English verbs divide into two broad as- pectual classes, my terms for which are perfective and imperfective.’’ The distinction is of a ‘‘primal character,’’ since it is related to the basic cognitive capacity of per- ceiving change (perfective) or the lack thereof (imperfective) (Langacker 1987: 258). As a diagnostic for classifying a verb as either perfective or imperfective, Langacker uses its compatibility with the progressive form: only perfectives occur in the progressive. The same test was used by Vendler to distinguish states from nonstates. Consequently, Langacker’s perfective processes cannot be equated with the telic situation types from the traditional literature discussed above. More specifically, Vendler’s atelic activity verbs (walk, swim, sleep) denote perfective processes in Langacker’s approach (see Kochanska 2000: 144). While the ‘‘grammatical’’ terminology of perfective/imperfective aspect is ap- plied here to a Vendler-like lexical classification of verbs, Langacker does warn against treating these classes as a rigid lexical partitioning: verbs may have a default value, but the aspectual interpretation of a given expression is ‘‘flexibly and globally determined’’ (1999: 390, note 14). This is illustrated in (30). (30) a. The road winds through the mountains. b. The road is winding through the mountains. The verb wind in (30a) is imperfective; the sentence does not express change. The verb wind in (30b), however, is considered by Langacker to be a perfective verb since it is used in the progressive form. These sentences, therefore, nicely illustrate the crucial role of construal in the domain of aspect: one and the same situation may be construed as either perfective or imperfective. To complicate matters, the progressive itself, providing an internal perspective on a perfective process, is called an ‘‘imperfectivizing device,’’ making the process expressed by the sentence (30b) as a whole imperfective. As the discussion of (30) shows, Langacker does not make a fundamental distinction between lexical and grammatical aspect but rather subsumes both under the common concept of ‘‘perfectivity,’’ which is applied to verbs, constructions, as well as to complete expressions. Croft (1998) refers to Langacker’s perfective and imperfective processes as actions (involving change) and states (involving no change), respectively. Actions are subdivided into processes (extended in time) and achievements (not extended in tense and aspect 815 time). In addition, Croft introduces the notion of point state (no change, not extended in time) for such things as It is eight o’clock and The train is on time. Croft addresses the complex interplay between the temporal structure of events as named by verbs and ‘‘aspectual grammatical constructions,’’ such as the simple/progres- sive distinction in English. On the one hand, lexical aspect seems to determine certain grammatical patterns. This is evidenced, for instance, in the constraint on the use of the simple present tense in English demonstrated in (24b). On the other hand, aspectual constructions themselves provide a conceptualization of the tem- poral structure of the event, and languageusers are flexiblein adjusting the temporal structure to fit the construction. This is illustrated in (31) (Croft 1998: 71). (31) a. ?I am loving her. b. I am loving her more and more, the better I get to know her. The verb love is one of Vendler’s examples of states (see 26a), and, accordingly, it does not easily appear in the progressive form (see 31a). However, a shift in ‘‘temporal scale’’ leads to a shift in acceptability. In (31b), the state of loving turns out to involve change, and thus the progressive can be used. Examples such as these (and the sentences in 30) lead Croft to conclude that verbs cannot be divided into word classes on the basis of the grammatical constructions in which they occur. (He makes a similar point for the test involving adverbials as in 27 and 28). Croft, therefore, seriously questions the value of ‘‘distributional analysis’’: words do not have a fixed distribution across constructions; the interaction between lexicon and grammar is ‘‘mediated by conceptualization processes’’ (1998: 79). Contrary to Langacker and Croft, Michaelis (1998) sharply distinguishes be- tween lexical and grammatical aspect, as situation aspect and viewpoint aspect, re- spectively (following the terminology of Smith 1991). As for the former, she regards the distinction between events and states as a universal cognitive distinction, in- dependent of its specific manifestation in a given language. Viewpoint aspect, the perfective/imperfective distinction, concerns the grammatical encoding of the event/state distinction using language-specific resources. There is no one-to-one mapping between situation aspect and viewpoint aspect, since viewpoint aspect may override ‘‘the canonical representation’’ of situations, as was exemplified in (31) (see Smith’s 1991 marked and unmarked aspect choice). English, in Michaelis’s view, does not grammatically encode the event/state distinction at all, which makes viewpoint aspect a covert category of English. Michaelis departs from a long-standing tradition in the aspect literature by not treating the progressive as an expression of imperfective aspect, but introducing a third category in addition to situation aspect and viewpoint aspect, namely the category of phasal aspect. This is a cover term for (i) inceptive aspect (start, begin), (ii) the progressive, and (iii) the perfect. In Michaelis’s view, the progressive does not directly encode the event/state distinction, but rather presupposes it and ac- complishes a perspectival shift from an event predication to a state predication. Michaelis, therefore, calls the progressive an ‘‘override construction.’’ The same intuition was captured by Langacker in his treatment of the progressive as an 816 ronny boogaart and theo janssen ‘‘imperfectivizing‘‘ device. And, in his discussion of (31), Croft as well assumes that the ‘‘default’’ semantic class of a verb may be altered by the grammatical envi- ronment it occurs in (cf. the notions of ‘‘coercion’’ and ‘‘shift’’ discussed by Moens & Steedman 1988 and Hayase 1997). 3.3. Meaning The cognitive linguistic literature on the semantics of aspect departs from tradi- tional accounts in two ways. First, rather than defining aspectual categories by single semantic contrasts (e.g., complete/noncomplete, durative/nondurative), the semantics of aspectual categories is assumed to be organized around a prototype with many language-particular extensions, including extensions in other domains (tense, modality). Such an approach to aspect can already be found in the work of Hopper (1979) and has received wide support from large-scale typological studies (Dahl 1985) and from the grammaticalization literature (Bybee, Perkins, and Pagliuca 1994; Carey 1994). Case studies within Cognitive Grammar are offered by Midgette’s (1995) analysis of the progressive in Navajo, by Kochanska’s (2000) work on the semantics of aspect in Polish, and by Dickey’s (2000) comparative analysis of aspect in the Slavic languages generally. Second, in the cognitive linguistic literature on aspect, the focus of attention has shifted from defining aspect in terms of ‘‘the internal temporal constituency of a situation’’ (Comrie 1976: 3) to describing the function of aspectual distinctions at the discourse level. In particular, aspect is said to indicate viewpoint and to play a role in establishing relations across clauses. 3.3.1. Aspect and Viewpoint To describe the difference between the French perfective past passe ´ simple and the imperfective past imparfait,Cutrer(1994) uses the descriptive tools from mental space theory outlined in section 2.2. She defines the difference between perfective and imperfective aspect as a difference in perspective: the imperfective past indicates that the Focus space is also Viewpoint space, whereas the perfective past does not indicate viewpoint. The way Cutrer describes im perfective and perfective is shown in (32) and (33) and graphically represented in figures 31.3 and 31.4 (Cutrer 1994: 193–95). (32) The imperfective identifies a focus space N and indicates that N is viewpoint. (33) The perfective identifies a focus space N and indicates that N is not viewpoint. As was explained in 2.2.1, the past tense sets up a past Focus space in relation to a parent Viewpoint space. In fact, the representation for perfective past in figure 31.4 is identical to the one for past in figure 31.2; the perfective past does not indicate Viewpoint. The situation is thus construed from an external Viewpoint (in the parent space M). The imperfective past, however, indicates that the past tense and aspect 817 Focus space is also a Viewpoint. Thus, imperfective aspect establishes a shift from an ‘‘external’’ Viewpoint position to an ‘‘internal’’ one. (The ‘‘external’’ Viewpoint in space M, however, may still be available. Space M may, for instance, be a speech or thought space, carrying its own Viewpoint role, as in indirect speech.) Cutrer’s analysis of the perfective/imperfective distinction in terms of perspective is sup- ported by many observations regarding the use of imperfective forms in discourse representing the thought or speech of an individual other than the speaker, as in indirect speech and free indirect speech (Ehrlich 1987; Caenepeel 1989). 3.3.2. Aspect and Discourse Relations Aspectual information also turns out to be relevant for determining relations be- yond clausal boundaries. In fact, according to Hopper (1982: 5), ‘‘The fundamental notion of aspect is not a local-semantic one, but is discourse pragmatic.’’ Hopper argues that the distinction between perfective and imperfective aspect primarily serves to indicate foreground and background in discourse: The perfective aspect is found mainly in kinetic, sequential events which are central to the unfolding of the narrative. Imperfective aspect is used typically for backgrounding: situations, descriptions and actions which are simultaneous or overlapping with a perfective event. (Hopper 1979: 58) Thus, aspect is said to convey information on the temporal ordering of situations presented in consecutive sentences: perfective forms present sequential situations, making up a narrative chain of events; imperfective forms present situations that are ‘‘going on’’ in the background (see Fleischman 1985). Reinhart (1984) has shown that the distinction between foreground and background is analogous to the dis- tinction between Figure and Ground in visual perception, ‘‘neutral and clearly unavoidable organization systems’’ (790); this analogy was already noted in the writings of Talmy (1978) (see also Hayase 1997). While Hopper restricts his discussion to grammatical forms of (perfective/ imperfective) aspect in various languages, a highly similar approach has been Figure 31.3. past imperfective 818 ronny boogaart and theo janssen proposed for English within Discourse Representation Theory in terms of the dis- tinction between events and states (Kamp and Reyle 1993). In this framework, the Reichenbachian notion of reference point (see section 2) is used to formalize the intuition expressed above by Hopper. Tenses in discourse are always interpreted with respect to a reference point provided by a preceding sentence, but while events follow their reference point, states are said to include their reference point (Kamp and Reyle 1993: 528). Thus, the situation of John walking to the bookcase in (34a) is correctly predicted to follow the situation of John opening the door, which serves as reference time. Together they form an iconically ordered sequence of events (i.e., Hopper’s ‘‘foreground’’). (34) a. John opened the door and walked to the bookcase. b. John opened the door. It was pitch dark in the room. The second sentence of (34b), however, presents a state, which includes the ref- erence point provided by the previous sentence: it was dark both before and after John opened the door (i.e., Hopper’s ‘‘background’’). While it seems clear that aspectual information has a role to play in the process of interpreting temporal relations in discourse, it is not possible to maintain that aspect alone determines temporal relations. It should be noted, for instance, that the linguistic information provided by (34b) does not exclude the possibility that the room was not dark at all before John opened the door. In fact, the pre- ferred interpretation of (35) is one in which the state of the room being dark follows but does not precede the event of the first sentence. (35) John switched off the light. It was pitch dark in the room because the Venetian blinds were closed. (Hinrichs 1986: 68) Interestingly, Cutrer’s ‘‘perspective approach’’ to aspect, mentioned in section 3.3 .1, does seem applicable to both (34b) and (35): the second sentence describes the state of the room being dark from John’s perspective, in other words, as the first thing he noticed after opening the door or switching off the light. To explain the Figure 31.4. past perfective tense and aspect 819 . to the usefulness of classifying verbs as such. The problem is all the more urgent since the influence of other elements in the clause is not restricted to the direct object. The nature of the. one in which the state of the room being dark follows but does not precede the event of the first sentence. (35) John switched off the light. It was pitch dark in the room because the Venetian. implied by the use of English verbs’’ (144). He divided the verbs of English into four classes depending on (i) whether or not the situation as expressed by the verb has duration, (ii) whether the situation involves