Frančiška Lipovšek 186 different prepositions can be used for one and the same constellation of objects in the real world, which is due to different possible conceptualizations (cf. Šarič 2001, 2006). Last but not least, prepositions display a great deal of crosslinguistic variation because languages vary in their semantic construal of space (cf. Talmy 2000: ch. 3.2.2). All this speaks against the classical approach to the semantics of prepositions. Lakoff (1987: 418-61) understands prepositions as a radial category with a central image schema from which all the other uses are derived. Brala (2007: 4) speaks of a dichotomy between “cognitive universality and “linguistic relativity” and distinguishes between the “(deep) level (of conceptual universals)” and the “surface level of lexical patterns” (Brala 2007: 25). Coventry (1998: 260) proposes an interface between the spatial world and language in the form of “mental models.” Both argue that spatial prepositions should be understood not only in terms of geometric relations, but primarily in terms of functional relations based on interaction (cf. also Talmy 1988). Brala (2005, 2007) understands this interaction in terms of control: the TR’s location is controlled through containment or support and is conditioned by dynamic factors such as dimensionality, orientation and attachment. The same view is held by Cuyckens (1988), who describes the conceptual structure of in in terms of (functional) containment rather than (geometric) inclusion, arguing that only a container can control the position of the contained. The present paper focuses on three prepositions expressing the TR’s movement away from the LM: out of, off and from. It can be assumed that each of these prepositions triggers an abstract schema reflecting the TR- LM configuration in the static scene serving as the departure point. While out of and off are relatively straightforward in this respect, from is rather special because it is compatible not only with point-like LMs but also with the TR-LM configurations evoked by the other two prepositions. A short comparison with Slovene is included in the paper to make this universality of from even more evident. The aim of the paper is to find out whether from plays any role in the conceptualization of the departure point and how that affects its interchangeability with out of and off. 1. Static scenes serving as departure points For any movement that is clearly bounded by a departure point, this point is conceptualized in terms of a static scene structured around some kind of functional relation between the TR and LM. For example: (1.) An apple rolled out of the bag. (2.) She accidentally knocked the vase off the table. (3.) The child toddled from the table to the sofa. Movement out of , off , and from the Landmark 187 For (1.), the departure point is conceived of as a relation between a threedimensional container (LM) and an object (TR) within the container. The LM performs the function of containment. The container schema of (1.) can be evoked by spatial in: The apple was in the bag. For (2.), the departure point is conceived of as a scene where one object (TR) is in contact with and supported by another object (LM). The LM provides support for the TR. The support schema of (2.) can be evoked by spatial on: The vase was on the table. For (3.), the departure point is conceived of as a scene where one object (TR) is in proximity to another object (LM). The function of the LM with regard to the TR is not specified. The proximity schema of (3.) can be evoked by spatial by: The child was standing by the table. The container and support schemata share the function of control. In the former, the location of the TR is controlled through containment: the walls of the container prevent the TR from falling out. In the latter, the LM controls the TR through vertical support. 2 No control function, by contrast, can be recognized in the proximity schema. The LM in (3.) serves merely as an orientation point for locating the TR, but there is no interaction between the two. If the mental representations of movement in (1-3) derive from the prepositions used, and if the conceptualization of each movement involves the conceptualization of its departure point, it can be assumed that dynamic out of evokes the same container schema as static in, dynamic off the same support schema as static on, and dynamic from the same proximity schema as static by. This, of course, does not mean that there is a one-to-one match between these prepositions. The container schema can, for example, be evoked by inside, the support schema by atop, and the proximity schema by next to: 3 (4.) a. What’s inside the box? b. Let’s take it out of the box. 2 That, however, suffices only if the LM, or rather, the surface enabling contact with the TR is positioned horizontally. If the case of a vertical position or a tilt, the TR has to be attached (i.e. fixed in position) to the LM in order to counteract gravity. 3 Inside, atop and next to are more specific in meaning than in, on and by. Inside seems to profile position (i.e. the TR is conceived of as located within the walls of the container) rather than containment itself (i.e. the TR is conceived of as (part of) the contents of the container). The mental representation of an atop-scene involves an object with a clearly recognizable top part. Next to profiles the nontouching dimension of proximity: (1.) *The syrup/ The energy rating label must be somewhere inside the fridge. (2.) She was sitting *atop/ on her bike. (3.) He stood *next to/ by the door, leaning against the doorpost. Frančiška Lipovšek 188 (5.) a. She’s sitting atop a two-metre high wall. b. If she fell off that wall, she’d kill herself. (6.) a. He was standing next to the window. b. He rushed from the window to the door. The mental representation of movement (in a direction) away from the LM is a combination of two different schemata triggered by the preposition: (i.) a container, support or proximity schema representing the departure point and (ii.) a path schema representing the movement itself. The path schema presents the TR as moving in a direction away from the LM. The path begins at a point whose location with respect to the LM is determined by the static component of the representation. The major difference between the uses of out of, off and from lies in the conceptualization of the departure point. With out of, the movement typically begins inside the LM. The TR reaches the boundaries of the LM and exits the LM. Whether or not it moves further away is irrelevant: (7.) Take that chewing gum out of your mouth! (8.) She stormed out of the room. With off, the movement typically begins on the outside of the LM. It is possible for the TR to move across the (supporting) surface before leaving the LM. Whether or not it moves further away is irrelevant: (9.) She wiped the crumbs off the table. (10.) Someone has taken the mirror off the wall. With from, the movement begins at a point whose exact location with regard to the LM is irrelevant because the LM itself is perceived as that very point. From triggers a proximity schema ignoring the details of the spatial arrangement: (11.) He rushed from the window to the door. (= 6b) (12.) We swam from one end of the pool to the other. It is completely irrelevant in (11.) whether the man, before rushing to the door, sat in a rocking chair half a metre away from the window or stood by the window, leaning on the windowsill. Whichever the man’s microlocation, his proximity to the window makes it possible to perceive the window as the starting point on his path. What is relevant for further discussion is the observation that in both sentences the initial TR-LM relationship is that of external proximity: the TR is located outside the LM. There may be direct physical contact between the two, but the function of the LM remains to be that of a mere Movement out of , off , and from the Landmark 189 orientation point. The proximity schemata recognized in (11.) and (12.) can be evoked by static by and at respectively. 4 2. The role of from in conceptualization From is special in that it is compatible with non-dimensional as well as dimensional LMs, in which case the TR can be located not only in (external) proximity to the LM but also internally (cf. Lindstromberg 1998: 45). The role of from thus deserves careful attention in cases where the initial TR-LM relationship is clearly that of support or containment: (13.) Could you clear your things from the sofa? (14.) He took a hammer from the toolbox. Does from in (13.) and (14.) still trigger a proximity schema for the departure point? It can easily be replaced by off and out of respectively: (15.) Could you clear your things off the sofa? (16.) He took a hammer out of the toolbox. The dimensionality of the LM seems less relevant in cases like: (17.) Could you bring me a pillow from/ *off the sofa, please? (18.) Could you bring me the hammer from/ *out of the toolbox, please? But even in (17.) and (18.) the initial (i.e. pre-movement) relationship between the TR and LM is conceptualized in terms of the support and container schemata respectively: a pillow on the sofa; the hammer in the toolbox. Nevertheless, the two are not triggered by from but are evoked on the basis of the default relationship between the TR and LM: it is reasonable to expect that there will be pillows on sofas and hammers in toolboxes. The same holds for (13.) and (14.) above. If the TR and LM are such that they allow different geometric configurations between them, ambiguity may arise: (19.) All of a sudden, a cat jumped from one of the boxes. Was the cat standing on the box and jumped off it or was it in the box and jumped out of it? It can be either. Whichever the case, sentence (19.) does not say anything about the initial relationship between the TR and LM - this must be either already known or simply irrelevant. With off and out of, by contrast, that very relationship is placed in the foreground: 4 It should be noted that by triggers an external proximity schema while with at the actual TR-LM arrangement is irrelevant because the LM is conceived of as a geometric point coincidental with the TR (cf. Cienki 1989: ch. 3.2.1, Keizer 2008, Knaś 2006). The two prepositions differ also in the functional component: at is used when the TR is oriented towards the functional side of the LM (e.g. sitting at the computer), which is absent in by (e.g. standing by the car) (cf. Knaś 2006). Frančiška Lipovšek 190 (20.) Something startled the cat and it jumped off the box. (21.) The cat managed to jump out of the box and escape. An immediate question arising at this point is, of course, that of interchangeability. In sentences (17.) and (18.) above, repeated below as (22.) and (23.) respectively, the only option is from: (22.) Could you bring me a pillow from/ *off the sofa, please? (23.) Could you bring me the hammer from/ *out of the toolbox, please? The conceptualization of bring involves a path whose starting point is a point-like LM: the object moves from point A to point B. That is incompatible with off and out of, which profile a path that begins somewhere within the boundaries of the LM and ends at the point where the TR leaves the LM. With containers as LMs, the TR necessarily covers a certain path already before exiting the LM. With surfaces, by contrast, that is not necessarily the case. In (24.) below, the TR moves across the surface of the LM already before leaving it, while in (25.) the TR’s movement begins with its very detaching itself from the LM: (24.) The child rolled the teddy bear off the sofa and brought it to me. (25.) The child lifted the teddy bear off the sofa and brought it to me. Whichever the case, the path schema associated with bring ignores the dimensionality of the LM and cannot include any movement implied by the support and container schemata. Any movement prior to the TR leaving the LM is conceptualized as part of another event, which is reflected in the semantic implications of (22.) and (23.) above respectively: (26.) Could you take a pillow from/ off the sofa and bring it to me, please? (cf. (22.)) (27.) Could you take the hammer from/ out of the toolbox and bring it to me, please? (cf. (23.)) With bring, the TR’s path begins with a point-like LM and leads to another point-like LM. The point-like LM is part of the departure-point proximity schema activated by from. It can be assumed that whenever the TR’s path is necessarily conceived of as a line between two points, from is required. In (28.) below, for example, it is possible to foreground the relationship that existed between the TR and LM before movement. In (29.), by contrast, the combination with to activates a path schema that profiles a line between two points, which makes from the strongly preferred option. (28.) a. The cat jumped from/ off the box. b. The apple rolled from/ out of the bag. Movement out of , off , and from the Landmark 191 (29.) a. The cat jumped from the box to the windowsill. b. The apple rolled from the bag to the other side of the table. It is nevertheless possible to justify the use of off and out of on the grounds that the whole event is perceived in terms of two separate phases. In the first phase the TR crosses the boundaries of the LM, which means that it necessarily reaches a point outside the LM. In the second phase the TR moves from that very point to another point on its path: (30.) a. The cat jumped off the box and (from there) to the windowsill. b. The apple rolled out of the bag and (from there) to the other side of the table. If we now return to the question addressed in connection with sentences (13.) and (14.) above, i.e. whether from triggers a proximity schema for the departure point also in cases where the LM is clearly conceived of as a surface or container, the answer is affirmative. Moreover, it is exactly this characteristic that makes from the only option in prepositional phrases expressing source: (31.) a. Someone had torn several pages out of her diary. b. He found several pages from/ *out of her diary in the garbage. (32.) a. Has anyone taken the directory off my desk? b. This directory will be from/ *off your desk, I suppose. As illustrated by these examples, only from is possible in (31b) and (32b) although the initial relationships between the TR and LM do not differ from those in (31a) and (32a) respectively: the pages were in the diary; the directory was on my desk. But as the LM represents a source, it is not conceived of as a container or surface, but rather as the starting point on the TR’s path: several pages (that came) from her diary; a directory (that comes) from your desk. That is why out of and off need to be replaced by from. The TR moves away from a point-like LM, meaning that the departure point is necessarily conceptualized in terms of a proximity schema. Nevertheless, the LM can function as source as well as location if the TR represents some kind of continuum, for example an aggregate of objects or a substance (which consists of smaller amounts or portions of various sizes and is thus comparable to an aggregate of objects). An object may be taken out of its original location, but will still represent the same aggregate of objects. And as long as there is at least one object remaining in the original location, the relationship between the TR and LM can be perceived in either way. It is crucial to note, though, that although Frančiška Lipovšek 192 on and in are possible, off and out of remain unacceptable because they are incompatible with the point-like character of the source. (33.) The candies from/ on/ *off that plate are sweeter. (34.) The water from/ in/ *out of that fountain is not drinkable. It can be concluded that the major feature that sets from apart from off and out of is that of reducing the LM’s dimensionality in space to a mere point on a line: (35.) I heard an angry voice coming from their table. Since the only relationship a point-like LM can enter with the TR is that of proximity, the preposition from cannot say anything about the initial functional relationship between the TR and LM: (36.) a. X fell out of Y. (← X was in Y.) b. X fell off Y. (← X was on Y.) c. X fell from Y. (← ? ? ? ) With from, the conceptualization of the departure point derives from the default relationship between the TR and LM. In (35.) above, the person whose voice I heard was probably sitting at the table. In (37.) below, the person was in the closet: (37.) I heard an angry voice coming from the closet. As pointed out by Keizer (2008), from can combine with other prepositions, but those expressing the default relationship between the TR and LM are redundant: (38.) I heard an angry voice coming from *at their table/ *in 5 the closet/ *on the balcony. An additional preposition is needed only if the initial functional relation between the TR and LM differs from the expected default one: (39.) The voice was coming from under the table/ behind the closet/ above the balcony. The conceptualization of the departure point now involves the second preposition. From still triggers a proximity schema (i.e. the TR’s path begins at a point represented by the LM), but the exact configuration of the TR in LM is part of a mental representation evoked by the second preposition. All this, however, should not imply that from suffices whenever the TR-LM relationship is a default one. The factor that needs to be considered is not the relationship as such but its relevance to the message: 5 In is redundant, but inside is not: I heard an angry voice coming from inside the closet. This supports the view that inside is more specific in meaning than in (see fn. 3). Movement out of , off , and from the Landmark 193 (40.) Take the butter from the fridge and cut it into slices. (41.) Take the butter out of the fridge and wait till it softens. The fact that butter is normally kept in a fridge is irrelevant in (40.). The message of the sentence would not change if the butter was placed elsewhere: (42.) Take the butter from the shopping bag/ the shelf and cut it into slices. The message of (40.) is clear: take the butter (from wherever you keep it) and cut it into slices. In (41.), by contrast, the container function of the fridge plays a significant role in the message. The butter will soften only at room temperature, i.e. outside the fridge. The location of the butter is necessarily some place cool enough to prevent it from softening. Nevertheless, in cases like (40.) and (41.) above, either preposition is theoretically possible. But out of is superfluous in (40.) because the exact functional relationship between the TR and LM is not important. It suffices if the LM is understood as the location of the TR, and from does the job perfectly well. In (41.), by contrast, the relationship between the TR and LM is crucial for the message. And the use of out of is simply more economical in this respect. It makes the relationship clear through the container schema it activates, while from calls for some extra effort on the part of the hearer, which involves identifying the default relationship between the TR and LM. As with concrete LMs, from triggers a point-like conceptualization also in the abstract domain: (43.) Peter was just coming out of the library. (44.) Peter was cycling from the library to the park. (45.) Peter has just come back from the library. In (43.), Peter was leaving the library building, which is necessarily conceptualized as a container. In (44.), the library building is conceptualized as the starting point on Peter’s path. In (45.), the library also represents the departure point, but it does not refer to the very place. The message of the sentence is that Peter has just returned from some activity that normally takes place in a library, namely reading and borrowing books. Instances like (45.) felicitously combine metonymy and metaphor. The name of a building or facility is used to refer to the activity taking place there, and this metonymic relationship calls for a metaphoric extension of the notion of departure point from a concrete location to an abstract activity (cf. Johnson/ Lakoff 2002, Lakoff 1987: 77, 276-78). Frančiška Lipovšek 194 (46.) a. We didn’t get back from the cinema until midnight. b. She came back from the supermarket with heavy bags. c. I’m just back from the pool and I’m starving. By contrast to (45.) and (46.), the dimensionality of the LM turns relevant in the following example: (47.) There’s something about swimming that reminds me of dancing, in that your mind gets in touch with your breathing, with your body in a rhythm. In a sense, you come out of the pool the way you come out of class, more centered, more relaxed, and you feel healthy, and there’s a lot of oxygen going through your body. (Teresa Reyes, cit. in Martins 1997: 51) A dance class, like any other activity, can be conceived of as a point on a line: one gets home from class in the same way one gets home from swimming. But in our example, coming out of class refers to the time immediately after the class is over, and is comparable to climbing out of the pool after swimming. The class is conceptualized as a threedimensional container filled up with physical activity. The metaphor could not be more obvious: one can immerse oneself in an activity in the same way one gets immersed in water. Coming out of class in (47.) refers to leaving an activity that is confined to certain spatial and temporal boundaries, i.e. the ‘here and now’ of the dance class in question. But involvement in an activity can be meant also in a very general sense. In (48.), for example, coming out of college refers to leaving college education in general: (48.) a. Most of the applicants have just come out of college. b. He dropped out of college. c. Her parents took her out of college. A change in conceptualization, however, whereby the activity is reduced to a point on a line, requires from: (49.) She moved smoothly from college into her dream job. 3. Out of, off and from in Slovene The prepositions out of and off are rendered into Slovene by iz and s/ z 6 respectively. From seems to have a direct counterpart in the preposition od: 6 The choice between (voiceless) s and (voiced) z depends on the voicing of the first sound of the following word: the former occurs before the letters p f t s c č š, the latter before all letters representing voiced sounds (cf. Toporišič 2000: 413). Movement out of , off , and from the Landmark 195 (50.) a. An apple rolled out of the bag. (= 1) b. Iz torbe se je skotalilo jabolko. 7 out of the bag (51.) a. Someone has taken the mirror off the wall. (= 10) b. Nekdo je s stene snel ogledalo. off the wall (52.) a. We swam from one end of the pool to the other. (= 12) b. Plavali smo od enega roba bazena do drugega. from one end of the pool Nevertheless, as will be shown below, the prepositions from and od can hardly be treated as equivalents. In the following two sentences, for example, od is not an option in Slovene although both prepositions are possible in English: (53.) a. He took a hammer out of/ from the toolbox. (= 14) b. Iz/ *Od škatle z orodjem je vzel kladivo. out of/ *from the toolbox (54.) a. Could you clear your things off/ from the zofa? (= 15) b. Bi lahko z/ *od zofe pospravil svoje stvari? off/ *from the sofa Furthermore, od is not possible even in cases where from is the only option: (55.) a. Could you bring me the hammer from/ *out of the toolbox, please? (= 18) b. Bi mi lahko, prosim, iz/ *od škatle z orodjem prinesel kladivo? out of/ *from the toolbox (56.) a. Could you bring me a pillow from/ *off the sofa, please? (= 17) b. Bi mi lahko, prosim, z/ *od zofe prinesel blazino? off/ *from the sofa 7 The sentences under (b.) are Slovene translations of those under (a.). For the sake of brevity, only the relevant prepositional uses are glossed out. Frančiška Lipovšek 196 The same occurs with prepositional phrases expressing source: (57.) a. He found several pages from/ *out her diary in the garbage. (= 31b) b. Več strani iz/ *od njenega dnevnika je našel v smeteh. out of/ *from her diary (58.) a. This directory will be from/ *off your desk, I suppose. (= 32b) b. Tale imenik bo pa s/ *od tvoje mize, kajne? off/ *from your desk The only exception is cases where the path is foregrounded to such an extent that the LM necessarily loses all its dimensionality. Compare: (59.) a. Vsak dan se vozim iz Maribora v Ljubljano. out of Maribor into Ljubljana (‘I travel daily from Maribor to Ljubljana.’) b. Koliko časa potrebuješ z vlakom od Maribora do Ljubljane? from Maribor to Ljubljana (‘How much time do you travel from Maribor to Ljubljana by train? ’) Nevertheless, if the dimensionality of the LM is retained, iz and s/ z seem perfectly compatible with the conceptualization of the TR’s path as a line between two points: (60.) a. Mačka je skočila iz škatle v košaro. (‘The cat jumped out of the box into the basket.’) b. Mačka je skočila s škatle na mizo. (‘The cat jumped off the box onto the table.’) c. Mačka je skočila iz škatle na mizo. (‘The cat jumped out of the box onto the table.’) d. Mačka je skočila s škatle v košaro. (‘The cat jumped off the box into the basket.’) What iz or s/ z foreground is internal location, i.e. the location of the TR within the boundaries of the LM: the cat was in or on the box before it jumped. The preposition od, by contrast, would imply that the cat stood by the box. The most likely interpretation of (61.) below is that the cat stood by the box, jumped and landed somewhere close to the table: (61.) Mačka je skočila od škatle do mize. (‘The cat jumped from the box to the table.’) Movement out of , off , and from the Landmark 197 The English version with from, however, is more likely to be interpreted in terms of a support schema as departure point, meaning that the cat stood on the box, jumped off it and landed on the table: (62.) The cat jumped from the box to the table. Od seems to evoke an external proximity schema with the TR outside the LM. From evokes a proximity schema in which the microlocation of the TR can in principle be of any kind. In short, the difference between the two prepositions is that od seems incompatible with an internal location of the TR: (63.) a. Take the butter from the fridge/ shelf and cut it into slices. (= 40) b. *Vzemi maslo od hladilnika/ police in ga nareži na rezine. from the fridge/ shelf In English, the LM’s function of container or support in cases like (63a) can be emphasized by choosing out of or off over from. In Slovene, the adverbs ven (‘out’) and dol (‘down’), reinforcing the meaning of iz and s/ z respectively, are occasionally used: (64.) a. Take the butter out of the fridge and wait till it softens. (= 41) b. Vzemi maslo ven iz hladilnika in počakaj, da se zmehča. out out of the fridge (65.) a. Be careful you don’t fall off the ladder! b. Pazi, da ne padeš dol z lestve! down off the ladder The difference between od and from is maintained also in the abstract domain. Compare: (66.) a. Peter was cycling from the library to the park. (= 44) b. Peter je kolesaril od knjižnice proti parku. from the library (67.) a. Peter has just come back from the library. (= 45) b. Peter se je ravnokar vrnil iz knjižnice. out of the library In both examples the library represents the starting point on Peter’s path, which calls for the use of from. But while in (66.) the departure point can be conceived only in terms of external proximity, the expected TR-LM relationship in (67.) activates the container schema: Peter was in the Frančiška Lipovšek 198 library. This difference explains the choice between od and iz in (66b) and (67b). The library in (67.) refers to the activity taking place in the library there rather than the place itself. As has already been pointed out above, abstract activities are easily conceptualized as containers. It is reasonable to expect that English from will be rendered into Slovene by iz: (68.) a. She came home from school/ hospital/ church/ the cinema/ the supermarket. b. Prišla je domov iz šole/ bolnišnice/ cerkve/ kina/ trgovine. out of school (etc.) Nevertheless, if the activity is referred to by the very name, iz is replaced by s/ z: (69.) Prišla je iz trgovine/ z nakupovanja/ / iz bolnišnice/ z zdravljenja. off shopping off (medical) treatment The use of z (‘off’) with shopping and treatment in (69.) suggests that the two activities are conceptualized in Slovene as two-dimensional surfaces. This is supported by the fact that when reference is made to attending an activity, the name of the activity normally takes na (‘on’) rather than v (‘in’): 8 (70.) Peter je na zdravljenju/ treningu/ tečaju/ angleščini/ plavanju/ klavirju. (‘Peter is on treatment/ practice/ course/ English/ swimming/ piano’) The question arising at this point concerns cases like (69.): why two different conceptualizations of one and the same activity? The answer probably lies in the fact that the container function of a supermarket, hospital, school, etc. as a place is so strongly present in our consciousness that the container schema is activated also when the nouns supermarket, hospital, school, 9 etc., are used in the abstract sense. Many activities conceptualized as two-dimensional surfaces in fact occupy slots in a person’s schedule and can as such be perceived as points on a time-line. That explains their compatibility with od: 8 Nevertheless, if the activity is meant in a general sense, it is conceptualized as a three-dimensional container: (4.) Nisem več v plavanju. (‘I’m not involved with swimming anymore.’) (5.) Sem čisto ven iz plavanja. (‘I’m completely out of swimming.’) 9 It should be noted that school combines with s/ z if it refers to a school as an institution. Compare: (6.) Iz šole prihaja pozno. (‘She comes back from school late.’) (7.) Mnogo naših študentov prihaja s te šole. (‘Many of our students come from that school.’) Movement out of , off , and from the Landmark 199 (71.) Pravkar sem prišla s klavirja/ od klavirja. off piano from piano (‘I’m just back from my piano lesson.’) The interchangeability of s/ z (‘off’) and od (‘from’) in such cases runs parallel to that of static na (‘on’) and pri (‘at’), as well as that of dynamic na (‘on’) and k (‘to’): (72.) Bila sem na klavirju/ pri klavirju. on piano at piano (‘I had a piano lesson.’) (73.) S klavirja/ Od klavirja grem naravnost na balet/ k baletu. off piano from piano on ballet to ballet (‘After the piano lesson, I go straight to my ballet class.’) It should be noted at this point that pri (‘at’) also triggers the external proximity schema (cf. Knaś 2006, Šarić 2006). Nevertheless, if the TR is in contact with the LM and the LM loses its dimensionality, the external proximity relationship turns into that of coincidence with the whole LM. 4. Final remarks The question permeating the paper has been that of from and its interchangeability with out of and off. From evokes the picture of a point-like LM, meaning that it ignores its dimensionality and the internal location of the TR, which are, by contrast, the key components of the schemata triggered by out of and off. A factor to be considered at this point is the relative length of the TR’s path. The greater the distance covered, the smaller the relative size of the LM, so that its dimensionality is eventually reduced to zero. Keizer (2008: 224, examples (74) from same page) illustrates the role of distance on the basis of at/ by contrast, pointing out that (74a) is “appropriate only when the speaker is at some distance from the supermarket - which, as a result, loses its dimensions.” (74.) a. John is at the supermarket. b. John is in the supermarket. Furthermore, if distance is foregrounded, less weight is placed on the initial TR-LM relationship, which, conversely, leaves more room for interpretation. To illustrate the point, while (75a) below implies that the squirrel jumped off one stump and landed on the other (with the squirrel standing on the stump as the expected departure-point scene), (75b) is open to another interpretation; if the question is about how far the person Frančiška Lipovšek 200 can jump, it may simply suggest jumping from a spot near one stump to a spot near the other. (75.) a. The squirrel jumped from this stump to that one. b. Can you jump from this stump to that one? What should also be taken into consideration is the path covered by the TR before crossing the boundaries of the LM. With the container schema as departure point, the TR has to reach the wall of the container in order to break through the surface and exit the LM. With the support schema, by contrast, the TR is already placed on the surface of the LM, so in principle no prior movement is required to get detached from the LM. For example, while the pencil in (76a) moves across the surface of the desk first to reach its edge, the directory in (76b) is simply picked up from the desk. (76.) a. The pencil rolled off the desk. b. Somebody took the directory off my desk. But it is crucial to note that in both (76a) and (76b) the TR is already on the surface of the LM, which never applies to containers as LMs. This difference could have implications for prepositional use and the degree of interchangeability. For example, a native speaker consulted for judgments on some of the examples labelled neither of the sentences below as perfectly acceptable, but still considered (77a) more acceptable than (77b): (77.) a. The cat jumped out of the box onto the windowsill. b. The cat jumped off the box onto the windowsill. In cases where the TR moves from one LM to another, from seems to be more readily interchangeable with out of than with off. The reason could be exactly that with containers as LMs the TR has to cover some distance, however short, before reaching the actual boundary of the LM. This, in a sense, rescues the LM’s dimensionality. 5. Conclusion The prepositions out of, off and from differ in the roles they play in the conceptualization of the departure point. While out of and off evoke the container and support schemata respectively, from triggers a proximity schema in which the whole LM, irrespective of its dimensions, is conceived of as a point on a line. As such it cannot contribute to the conceptualization of the actual TR-LM arrangement, which has to be inferred from the semantics of the TR and LM themselves. The degree of inter- Movement out of , off , and from the Landmark 201 changeability between from and out of or off depends largely on factors that can affect the perception of the LM’s dimensionality. The more relevant the TR-LM relationship in terms of containment or support, the less probable that the LM will lose its dimensions. Conversely, if the TR’s path is foregrounded, the LM’s dimensionality will be reduced to zero. There is, of course, a cline between the two extremes, and the nature of from allows both conceptualizations to be involved in a single case. The comparison with Slovene makes the special nature of from even more evident. It dispels the misconception that from has a full equivalent in od. In fact, it shows that in cases where from is theoretically interchangeable with out of or off, od is automatically ruled out because it is not compatible with internal location. The question of interchangeability does not arise in Slovene because od presupposes a conceptualization in which the TR-LM arrangement is either that of coincidence or that of external proximity. An exception can be found in the abstract domain: many activities that are basically conceptualized as surfaces occur with s/ z (‘off’) as well as od. Activities that are conceptualized as containers, by contrast, are found only with iz (‘out of’). This is not without implications for our conclusions. The fact that only a ‘two-dimensional’ activity can take od goes hand in hand with the observation that choosing out of over from is relatively more frequent a phenomenon than choosing off over from. Both suggest that the initial TR-LM relationship is more readily ignored with two-dimensional LMs - a speculation that calls for further research. References Brala, Marija. (2005). “Crosslinguistic diversities and cognitive universals in the (semantic) domain of SUPERPOSITION.” In: J. Granić (ed.). Semantika prirodnog jezika i metajezik semantike. HDPL 2004 Proceedings. Zagreb - Split. Brala, Marija. (2007). “Spatial ‘on’ - ‘in’ categories and their prepositional codings across languages. Universal constraints on language specificity.” In: Andrea Schalley/ Dietmar Zaefferer (eds). Ontolinguistics. How Ontological Status Shapes the Linguistic Coding of Concepts. Berlin & New York: Mouton de Gruyter. 299-329. Cienki, Alan (1989). Spatial Cognition and the Semantics of Prepositions in English, Polish, and Russian. Munich: Sagner. Coventry, Kenny (1998). “Spatial Prepositions, Functional Relations, and Lexical Specification.” In: Patrick Olivier/ Klaus-Peter Gapp (eds.) Representation and processing of spatial relations. Mahwah, NJ: Lawrence Erlbaum. 247-262. Cuyckens, Hubert (1988). “Spatial prepositions in cognitive semantics.” In: W. Hüllen/ R. Schulze (eds.) Understanding the lexicon. Tubingen: Max Niemeyer. 316-328. Evans, Vyvyan/ Andrea Tyler (2001). “Reconsidering prepositional polysemy networks: the case of over.” Language 77/ 4. 724-765. Goddard, Cliff (2002). “‘On’ and ‘on’: Verbal explications for a polysemic network.” Cognitive Linguistics 13/ 3. 277-294. Frančiška Lipovšek 202 Heine, Bernd (1997). Cognitive Foundations of Grammar. New York: OUP. Jackendoff, Ray (1983). Semantics and cognition. Cambridge, MA: MIT Press. Johnson, Mark/ George Lakoff (2002). “Why cognitive linguistics requires embodied realism.” Cognitive Linguistics 13/ 3. 245-263. Keizer, Evelien (2008). “English prepositions in functional Discourse Grammar.” Functions of Language 15/ 2. 216-256. Knaś, Iwona (2006). “Polish equivalents of spatial ‘at’.” Proceedings of the Third ACL-SIGSEM Workshop on Prepositions. Trento, Italy: Association for Computational Linguistics. 9-16. Lakoff, George (1987). Women, fire, and dangerous things. Chicago: U of Chicago P. Langacker, Ronald W. (1987). Foundations of Cognitive Grammar, Vol. 1: Theoretical Prerequisites. Stanford: Stanford University Press. Lindstromberg, Seth (1998). English prepositions explained. Amsterdam: Benjamins. Martins, Peter (1997). New York City Ballet workout: fifty stretches and exercises anyone can do for a strong, graceful, and sculpted body. New York: Quill/ William Morrow and Company. Šarić, Ljiljana (2001). “Prepositional categories and prototypes: contrasting some Russian, Slovenian, Croatian and Polish examples.” Glossos 1. 1-20. Šarić, Ljiljana (2006). “On the meaning and prototype of the preposition pri and the locative case: a comparative study of Slavic usage with emphasis on Croatian.” Razprave Instituta za hrvatski jezik i jezikoslovlje 32. 225-248. Talmy, Leonard (1983). “How language structures space.” In: H.L. Pick/ L.P. Acredelo (eds.). Spatial orientation: Theory, research and application. New York: Plenum. 225-282. Talmy, Leonard (1988). “Force Dynamics in Language and Cognition.” Cognitive Science 12. 49-100. Talmy, Leonard (2000). Toward a Cognitive Semantics. Cambridge, MA: MIT Press. Toporišič, Jože (2000). Slovenska slovnica. Ljubljana: Založba Obzorja. Frančiška Lipovšek University of Ljubljana, Faculty of Arts Ljubljana, Slovenia