The cartographic syntax of Lai in Mandarin Chinese

3.1.1 Function of lai ₁

For lai ₁ , it can be classified into two types. In the first type, it has physical spatial reading, implying the agent moving from one position to another position as shown in (19). In the second type, lai ₁ has no physical spatial reading as shown in (17).

Some researchers distinguish between the non-spatial and spatial reading of lai ₁ , treating them as two distinct words with separate semantic functions (Zhu 1982, Xing 1998, Cao 2002, Lu 2006, Wu 2023, Mariko 2010). In contrast, other researchers argue that both readings of lai1 should be analyzed as a single lexical or functional item, maintaining a unified representation despite their interpretational differences (Zhang and Wan 2015, Zhang 2003, 2020). Below, we will follow the first opinion, taking spatial lai ₁ and non-spatial lai ₁ as two different words with the same pronunciation.

It is assumed in this research that lai ₁ should be treated as two different words for their different syntactic distribution. First, non-spatial lai ₁ cannot co-occur with the negative marker meiyou ‘not’, but spatial lai ₁ can, as in (24). Second, location NP can be located after spatial lai ₁ , but for non-spatial lai ₁ , it cannot, as in (25). Third, non-spatial lai ₁ cannot co-occur with Aspect Marker le, but spatial lai ₁ can, as shown in (26).

For non-spatial lai ₁ ’s semantic function, as stated above, many researchers have explored it. Zhang and Wan (2015) don’t differentiate between spatial lai ₁ and non-spatial lai ₁ and argue that lái ₁ indicates displacement and purpose, where in some cases it represents physical displacement, as in (26a), while in other cases it signifies an abstract displacement in psychological space, bridging psychological distance, as in (26b).

Although Zhang and Wan (2015) analysis may capture some phenomena of lai ₁ in Mandarin, there is one major shortcoming. The lai ₁ may not necessarily be a displacement, indicating abstract or physical. Suppose that (26b) is uttered by a leader who is making a speech; in this scenario, the leader doesn’t use lai ₁ to bridge the psychological distance between him and the audience.

For spatial lai _1, we argue that spatial lai ₁ conveys physical displacement. Consider (27):

In (27), spatial lai ₁ indicates that the subject ta (he) is moving toward the speaker.

For non-spatial lai ₁ , Wang (2010) believes that it expresses the meaning of the near future. However, the near future meaning can be cancelled with the usage of the future modal verb hui ‘will’ as in (28).

In (32), there is no near-future implication. So non-spatial lai ₁ doesn’t indicate near-future.

Lu (2006) indicated that the main semantic function of lai ₁ , which loses its sense of displacement and function as a restrictive focus marker. Xiong (2013, 2015, 2017) proposes that restrictive focus is positioned outside TP. Tsai (2015) argues that the future modal hui (‘will’) occupies the head position of T.

Considering both analyses, if non-spatial lai ₁ functions as a restrictive focus marker, it would logically be positioned higher than the future modal hui (‘will’) and the subject. However, this arrangement contradicts observed language patterns, as illustrated in (28). Wu (2023) also proposed that non-spatial lai ₁ conveys a prospective aspect. We adopt Wei’s (2023) perspective in this article, assuming that non-spatial lai ₁ expresses a prospective aspect. Prospective aspect, as proposed by Comrie (1976), is an aspectual category that indicates a connection between the present state and a future event. While it functions similarly to the perfect aspect, which relates a past event to the present, the prospective aspect instead indicates a future event that holds present relevance, emphasizing an impending or anticipated action.

The different interpretations of spatial and non-spatial lai ₁ arise from their distinct syntactic positions, a point which will be further elaborated in Section 3. Maintaining that spatial lai ₁ denotes physical movement, while non-spatial lai ₁ indicates a prospective aspect.

3.1.2 Syntactic position of lai ₁

3.2.1 Functions of sentence-final particle lai ₂ and lai ₃

For lai ₃ , it occurs in the lai ₁ +VP+lai ₃ structure, as shown in (44). Zhang and Wan (2015) argues that lai ₃ in this structure functions as a directional verb, implying physical displacement. A directional verb cannot be omitted from the sentence; otherwise, the entire displacement event would be incomplete, as illustrated in example (45). However, in the lai ₁ +VP+lai ₃ structure, as shown in example (46), the deletion of lai ₃ does not affect the legality of the sentence, and the overall expression is not significantly impacted. Furthermore, as Wu (2023) argues, the main verb can be affixed with an aspectual marker, but lai ₃ cannot. Therefore, lai ₃ should not be regarded as a lexical verb.

Xu (2005) argues that lai ₃ functions as an aspect marker. However, lai ₃ behaves differently from typical aspect markers.

First, omitting lai ₃ does not significantly alter the meaning of a sentence, whereas omitting aspect-marking sentence-final particles such as le changes the meaning, as illustrated in (47) and (48):

Second, sentence-final aspect markers can co-occur with SFPs that express speaker attitude, whereas lai ₃ cannot, as demonstrated in (49):

Xu (1983) and other researchers have analyzed lai ₃ in this structure as a sentence-final particle.

For lai ₂ , it can occur simply after VP and appear in the sentence-final position, as shown in (50). Some researchers treat lai ₂ as an aspect marker. For example, Jian (2001) consider it a perfective aspect particle. However, there are some issues with this viewpoint. One piece of evidence against the claim that lai ₂ functions as an aspect marker is its compatibility with imperative sentences, as illustrated in (51). Additionally, lai ₂ can co-occur with adverbs that indicate gradual progression, such as manman (‘slowly’), as shown in (52). Notably, imperative sentences and adverb manman (‘slowly’) typically do not co-exist with perfective aspect markers, as evidenced by the ungrammaticality of (53) and (54). Since lai ₂ is compatible with these constructions, it cannot be classified as an aspect marker.

Moreover, Wu (2023) argues that lai ₂ without a spatial reading is a verb and is part of a non-continuous directional verb compound. However, as Xu (1983) mentioned, describing lai ₂ as part of non-continuous directional verbs does not comprehensively capture the linguistic phenomenon. Zongli and Min (2022) consider lai ₂ after VP to be a viewpoint aspect marker; however, this analysis cannot explain the displacement orientation conveyed by lai ₂ .

As mentioned before, based on the analyses of most researchers (Chao 1968, Tang 1979, Xu 1983) regarding the nature of lai₂’s and lai₃’s, it is assumed that both are sentence-final particles for the following reasons.

Hancil et al. (2015) define the basic properties of SFPs as follows:

Formally, SFPs are typically non-inflecting, monomorphemic units that are prosodically integrated with a host unit, receiving low-key intonation, lacking constituent status, and being positionally fixed at the end of that unit. They have no conceptual meaning, cannot be questioned or focused, and cannot be used as an utterance on their own; instead, they require a host unit. Functionally, SFPs fulfil various tasks related to discourse structure, speaker attitude, illocutionary force, and turn-taking. Structurally, the use of SFPs is not licensed by syntactic rules but is determined by the context, as they relate an utterance to various aspects of the communicative context, such as the speaker’s stance, the speaker’s illocutionary goal, and discourse cohesion (Hancil et al. 2015, 16).

First, lai ₂ and lai ₃ meet most of the properties proposed by Hancil et al. (2015). The word ‘typically’ indicates that not all SFPs display all of these features, as some SFPs in Japanese do not exhibit all these features (e.g., SFP ne forms a turn-utterance on its own). Therefore, SFPs are best understood as a gradable notion, whereby the concept of SFP-ness is specified along a continuum.

Second, although lai ₂ and lai ₃ do not satisfy the formal criterion of being devoid of conceptual meaning, many SFPs in Cantonese also imply conceptual meaning. For example, sin indicates ‘first’, gamzai means ‘almost’ (Tang 2020), zaa3 means ‘only’, and tim₁ means ‘also’ (Law 2004). Additionally, some linguists have argued that SFPs have consistent, context-independent meanings as in previous studies (Fung 2000, Sybesma and Li 2007).

Third, lai ₂ and lai ₃ are non-inflecting, monomorphemic units, receiving low-key intonation.

Fourth, lai ₂ and lai ₃ lack constituent status. Based on the question test, no question can be formed with the answer being lai ₂ and lai ₃ . According to the movement test, lai ₂ and lai ₃ cannot be moved to other positions in the sentence without rendering the sentence ungrammatical.

Finally, lai ₂ and lai ₃ are also syntactically ‘transparent’ (Hooper 1975) to tag questions, negation, and other syntactic processes as shown in (55) and (56):

These criteria suggest that lai ₂ and lai ₃ are situated close to the other prototypical SFPs.

Moreover, the fact that lai ₂ and lai ₃ occupy the most peripheral position within the utterance, one which may only be followed by vocatives and other sentence-final particles, substantiates their status as SFP.

So, these criteria suggest that lai ₂ and lai ₃ are situated close to the other prototypical SFPs.

We propose that Lai ₃ functions as an independent focus marker and simultaneously indicates displacement toward the speaker for the following reasons.

For indicating displacement function, since lai ₃ can only occur in sentences with displacement meaning and for sentences without displacement meaning, it is ungrammatical with lai ₃ as (57), (58), and (59) show. For (57), it can convey two meanings, with physical displacement and without physical displacement. And both are grammatical. For (58), it can only convey the non-physical displacement meaning. So, when lai ₃ is added to the end of the sentence, it becomes ungrammatical as (59) shows.

Thus, lai ₃ in this construction has a displacement reading.

For focus marking, Yuan (2003a) defines focus as the semantically prominent part of a sentence that the speaker intends the listener to pay special attention to. Regarding information focus, Halliday (1967) states that focus reflects new information. Yuan (2003a, b) categorizes sentences in which the entire sentence conveys new information as broad focus. Since lai ₃ is used to direct the addressee’s attention to the information denoted by the clause, as shown in (60), it is analyzed as a sentence-final particle of the focus type.

In (60), the presence of lai ₃ in (60a) conveys an emphatic tone that is absent in (60b). lai ₃ indicates that the entire clause presents information that the addressee should focus on. Since spatial lai ₁ also expresses displacement, the repetition of the displacement reading serves to reinforce emphasis, highlighting the importance of this information.

Additionally, we can test whether lai ₃ in lai ₁ +VP+lai ₃ construction marks new information by examining its compatibility with subsequent negative sentences, considering (61) and (62):

The grammaticality of (61) demonstrates that lai ₃ is compatible with the new information. The unacceptability of (62) suggests that lai ₃ contradicts presupposed information, further confirming its role as a focus marker.

Functioning to draw the attention of the addressee to the new information, lai ₃ falls into the category of the focus marker.

Since lai ₃ functions to direct the addressee’s attention to new information, it falls within the category of focus markers. We propose that the focus-marking function of lai ₃ results from the grammaticalization of the motion verb lai (‘come’).

From a phonetic perspective, lai ₃ [lə] is the result of phonetic weakening and monophthongization of the directional verb lai (‘come’).

From a syntactic perspective, its position at the end of the sentence increases the likelihood of further grammaticalization into a higher-level grammatical category. According to Hengeveld and Olbertz (2012), grammaticalization typically follows a pathway from the lower layer to the higher structural layers. This means that the displacement verb lai (‘come’), originally in the lexical domain, can grammaticalize into a focus marker by moving upward. Focus pertains to the CP domain (see Section 5 for further discussion).

From a semantic perspective, lai ₃ retains part of its original directional meaning. The fundamental reason why the directional verb lai is moved to the end of the sentence and used as a focus marker lies in its inherent semantic compatibility with focus functions. When lai is used as a verb of motion, its pragmatic function is to instruct the listener to move in a proximal direction toward the speaker, using the speaker as the reference point. In terms of communicative intent, this imperative nature establishes a link to its focus-marking function.

When movement is conceptualized in discourse, lai ₃ shifts the listener’s focus from their current state of attention to the upcoming speech act of the speaker. From the speaker’s perspective, lai ₃ serves as a focus marker by prompting the addressee to shift from a state of inattentiveness to one of attentiveness toward the information conveyed by the clause. This focus-marking function aligns internally with the movement function expressed by lai (‘come’) as a directional verb.

Thus, we hypothesize the following grammaticalization pathway for lai ₃ in lai ₁ +VP+lai ₃ construction:

Focus lai ₃ < displacement lai (‘come’).

For lai ₂ , many researchers classify it as a directional complement, indicating directional movement. However, it will be argued that directional complement and lai ₂ should behave differently.

First, the directional complement lai can appear with the perfective aspect verbal suffix -le in (63), while the sentence-final lai ₂ cannot in (64). It is assumed that there is a matching or agreement requirement on the lai ₂ and the perfective aspect verbal suffix -le. Such a requirement is, however, not observed in directional complement lai.

Second, lai ₂ can rescue the incompleteness effect of sentences as in (65), but directional complement lai cannot as shown in (66). Xiong (2020) claims that Chinese sentences always show the incompleteness effect, but it can be rescued by the realization of one of the functional categories above the argument structural layer. So lai ₂ ’s completeness function proves it is a functional category above the argument structural layer but not a directional complement.

As shown by the contrast between (65a) and (65b), lai ₂ can rescue the incompleteness of (65a). With lai ₂ , (65b) is a complete sentence.

As shown in (66), with standard directional complement lai, this sentence is still incomplete. So, the standard directional complements lai and lai ₂ belong to different categories and have different functions.

For the semantic function of lai ₂ , we will follow Fan (2014) analysis, stating that it indicates movement toward the speaker physically or abstractly.

When lai ₂ indicates physical movement toward the speaker, it signifies that something or someone is moving in the speaker’s direction, as shown in (67)

In (67), lai ₂ conveys that the subject ta (‘he’) physically moves toward the speaker.

When lai ₂ indicates abstract displacement toward the speaker, it emphasizes the path of event change rather than physical movement. In this case, the source refers to the subject’s state before the change, while the goal represents the subject’s state after the change is complete. lai ₂ marks the transition of an event from its initial state to its result state, with the speaker conceptually positioned at the endpoint, serving as the deictic center, and observing the entire process in retrospect. Here, movement occurs within the mental domain, where the source and goal are not abstract, but rather concrete. Consider (68):

In (68), the starting point of the event is ‘not singing’, while the endpoint is ‘singing’. The speaker is positioned at the endpoint of the event, observing the process unfold.

Therefore, in this context, lai ₂ here indicates the event progression from the starting point to the endpoint, without direct physical displacement.

In summary, lai ₂ consistently indicates movement, whether physical or abstract, with the deictic center – typically the speaker’s physical location or the event’s result state – serving as the reference point. The speaker conceptually positions themselves at the endpoint of the transition, observing the event’s progression.

Additionally, in some cases, some researchers (Wu 2023, Zongli and Min 2022) have analyzed lai ₂ as an aspect marker. Consider (69) and (70):

The contrast between (69) and (70) indicates that (69) emphasizes the process and the change from a state of being anxious to a state of calming down, while (70) emphasizes the state of calming the mind itself. In (71), lai ₂ indicates the change from the state of not singing to the state of singing. Overall, lai ₂ does not indicate physical movement, but rather a transition of the event from one state to another. This can be understood as an abstract displacement, where the source is the subject’s status before change and the goal is the subject’s status when the change is complete. This represents a metonymic semantic shift from the physical displacement reading of the particle lai ₂ .

In conclusion, lai ₃ indicates displacement toward the speaker while simultaneously functioning as a focus marker. Its focus-marking function arises from the grammaticalization of the directional verb lai (‘come’). In contrast, lai ₂ exhibits a deictic function, marking movement toward the deictic center, which can be either the speaker’s physical location in cases of spatial motion or the event’s result state in instances of abstract displacement.

3.2.2 Syntactic positions of lai ₂ and lai ₃

In this section, we first analyze the syntactic position of lai ₃ ’s followed by the analysis of lai ₂ . Based on the above analysis, lai ₃ has two different functions, ranging from indicating displacement toward the speaker to serving as a focus marker.

The following represents the syntactic structure of the clausal periphery under the Cartographic Approach proposed in this article. This section first discusses the rationale behind this structure, followed by an explanation of how the syntactic positions of lai₂ and lai₃ are determined within this framework based on the ordering of particles.

The hierarchical structure can be divided into four layers: vP, TP, and Complementizer Phrase (CP).

For the CP layer, we adopt Rizzi’s (1997) analysis, which further divides CP from bottom to top into Finite, Focus, and Force. In Rizzi’s (1997) framework, topics appear both above and below the focus. However, since topic phrases do not have sentence-final particles, they are not included in this discussion. The S.Asp (sentential aspect) category is adopted from Pan’s (2019) analysis of Mandarin sentence-final particles.

The TP layer includes temporal particles, classified under the temporal category as defined by Tang (1998). Temp serves as a convenient label for Temporal and corresponds to the temporal properties of the clause, positioned at the periphery of TP.

Regarding the vP layer, Event is a projection at the periphery of the vP domain, associated with the eventualities and aspectual properties denoted by the predicate (Tang 1998).

Based on the previous analysis of the functions of lai ₃ , in the structure lai ₁ +VP+lai _3, lai ₃ functions as a focus marker. Therefore, it is reasonable to suggest that the SFP lai ₃ merges in the head position of the Focus Phrase.

However, under this analysis, one question remains unresolved, which is how to explain the displacement function of lai ₃ .

Considering the functions of lai ₃ , we propose that lai ₃ occupies the head position of the EventP. This is followed by head movement to the head position of FocusP, as shown in the following tree diagram (in the diagram, for convenience, the head movement is represented in one step).

[...[_FocusPFocus+lai _3i [_S.AspPS.Asp[_FinitePFiniteP[_TPSpec[_T’T[_EventPSpec[_Event’ t_i[_vPt_j

To derive the right word order, it is assumed that Focus is head-initial while vP in this diagram undergoes movement on the surface to a position somewhere between the subject and lai ₃ , along the lines in Kayne (1994).

Based on HMC (Head Movement Constraint), movement from one head position to another is only possible between a given head and the closest head that asymmetrically c-commands it. But for convenience, the structure above is simplified for expository purposes by not showing the middle movement steps. And the syntactic movement operation can explain the dual functions of lai ₃

lai ₃ ’s syntactic position can be proved by the non-occurrence of modal verbs like yao will and neng can and lai ₃ as shown in (73). For modal verbs to occupy the syntactic position between vP and ForceP, their occurrence would block the movement of lai ₃ from the vP layer to the CP layer.

Additionally, lai ₃ moves from Event to Focus because lai ₃ is endowed with the feature [Focus], and hence, for feature checking, attracts lai ₃ to move from Event to Focus. This process involves head-to-head feature checking (Cormack and Smith 1996).

For the movement of vP to the specifier position of FocusP, Pan syntacticians (2021) propose that under the minimalist framework, each SFP heads a phase and bears an EPP feature. Complement-to-specifier raising is required as a last resort to satisfy the EPP. So, the movement of vP to the specifier position of FocusP is to meet the EPP of FocusP.

For the syntactic position of lai, we propose that lai ₂ occupies the head position of EventP for the following reasons.

First, as we mentioned earlier, Event is associated with the eventualities and aspectual properties denoted by the predicate and serves as a projection in the periphery of the vP domain. Since the deictic function of lai ₂ inherently ties it to the eventuality expressed by the predicate, as it marks the event’s directional development toward a deictic center, whether spatially or abstract. It is reasonable to place it in the head position of EventP. To derive the correct word order, vP also undergoes movement on the surface to a position between the subject and lai _2.

Second, the linear order between sentence-final particle mei and lai ₂ further supports this analysis. Lai ₂ appears in a position lower than mei, assuming lai ₂ occupies the head position of EventP and mei functions as an SFP of the temporal type in the head of TempP, which belongs to the inflectional layer (Tang 2022). This is illustrated in (74):

Third, analyzing the sentence-final lai ₂ as a sentence-final particle in the head position of EventP accounts for the particle cluster in (75), aligning with the syntactic hierarchy proposed in (72).

Example (75) further shows that the sentence-final lai ₂ can co-occur with ma, a sentence-final of the interrogative type, and le, a sentence-final particle in the head position of S.AspP, implying that syntactically, lai ₂ is in the lower position compared to le and ma.

Finally, as a sentence-final particle in the EventP position, lai ₂ cannot be modified by adverbials, consistent with other sentence-final particles in Mandarin, regardless of their semantics.

In conclusion, lai ₂ occurs in the head position of the EventP, situated between TP and vP. In contrast, lai ₃ is base-generated in the head position of EventP and subsequently moves to the head position of FocusP.