2.1Defining topic55.For a more detailed discussion on topic and focus in Korean, please refer to Park and Yeon (2023)Park, Chongwon, and Jaehoon Yeon 2023 “Information Structure in Korean: What’s New and What’s Old.” Journal of Pragmatics [Published online]. .

What we have in mind as our pivotal definition of topicality comes from Erteschik-Shir (2007)Erteschik-Shir, Nomi 2007 Information Structure: The Syntax-Discourse Interface. Oxford: Oxford University Press., as summarized in (5).66.The identifications of aboutness, referential givenness, and relational givenness are our own; Erteschik-Shir’s original work does not provide these identifications for (5).

While Erteschik-Shir (2007)Erteschik-Shir, Nomi 2007 Information Structure: The Syntax-Discourse Interface. Oxford: Oxford University Press. is not specific on the precise distinction between referential and relational givenness, we interpret her definition of givenness as encompassing both. Then, the topic must be defined as in (6).77.The definition presented in (6) is comparable to Gundel and Fretheim (2004)Gundel, Jeannette K., and Thorstein Fretheim 2004 “Topic and Focus.” In The Handbook of Pragmatics, ed. by Laurence R. Horn, and Gregory Ward, 175–196. Oxford: Blackwell Publishing,.

In (6a), relational givenness partitions conceptual representation of a sentence into two, [X Y], where X is what the sentence is about.88.This type of partition has been known by different names, such as presupposition-focus (Chomsky 1971Chomsky, Noam 1971 “Deep Structure, Surface Structure and Semantic Interpretation.” In Semantics: An Interdisciplinary Reader in Philosophy, Linguistics and Psychology, ed. by Danny D. Steinberg, and Leon A. Jakobovits, 183–216. Cambridge: Cambridge University Press.; Jackendoff 1972Jackendoff, Ray 1972 Semantic Interpretation in Generative Grammar. Cambridge, MA: MIT Press.), topic-comment (Gundel 1974Gundel, Jeannette K. 1974 “The Role of Topic and Comment in Linguistic Theory.” Ph.D. dissertation. University of Texas.), theme-rheme (Vallduví and Engdahl 1996Vallduví, Enric, and Elisabet Engdahl 1996 “The Linguistic Realization of Information Packaging.” Linguistics 34: 459–519. ), and topic-predicate (Erteschik-Shir 1997Erteschik-Shir, Nomi 1997 The Dynamics of Focus Structure. Cambridge: Cambridge University Press.). The relational givenness indicated by this type of partition, however, is independent of referential givenness.

Another issue we need to clarify is whether every sentence needs a topic. Some scholars, such as Jun (2019)Jun, Youngchul 2019 “hankwuke cengpokwuphyoci un/nun-uy uymi [A Semantics for (n)un as an Information Structure Marker: With a Focus on the Meaning of Contrast].” Eoneohag: Journal of the Linguistics Society of Korea 85: 99–137. and Choi (2016)Choi, Yoonji 2016 “hankwuke cengpokwuco yenkwu [A Study on the Information Structure in Korean Language].” Ph.D. dissertation. Seoul National University., argue that a topic is optional at least in Korean, based on common examples like (7).

Examples like (7) are widely observed in English as well, as in (8).

In both (7) and (8), the utterances do not include overt topics. However, it does not mean that there is no topic; rather, the topics are implicitly expressed as stage topics, which indicate the spatio-temporal parameters of the sentence (see Erteschik-Shir 2007Erteschik-Shir, Nomi 2007 Information Structure: The Syntax-Discourse Interface. Oxford: Oxford University Press., 16). We adopt Erteschik-Shir’s stage topic because our definition is based on Strawson (1964)Strawson, Peter F. 1964 “Identifying Reference and Truth-values.” Theoria 30: 86–99. , in which topics are the pivots for assessment; therefore, every sentence needs a topic.

2.2Reference point

Another concept related to topicality is reference point, as proposed by Langacker (1993Langacker, Ronald W. 1993 “Reference Point Constructions.” Cognitive Linguistics 4, 1–38. , 2001Langacker, Ronald W. 2001 “Topic, Subject, and Possessor.” In A Cognitive Approach to the Verb, ed. by Hanne Gram Simonsen, and Rolf Theil Endresen, 11–48. Berlin: De Gruyter Mouton., 2008Langacker, Ronald W. 2008 Cognitive Grammar: A Basic Introduction. Oxford: Oxford University Press. , 2009Langacker, Ronald W. 2009 Investigations in Cognitive Grammar. Berlin: Mouton de Gruyter. ). Humans have the basic cognitive ability to invoke the conception of one entity in order to make mental contact with another. The readily accessible entity with which the conceptualizer makes initial contact is called a reference point, and the less accessible entity contacted via the reference point is called a target. As a readily accessible entity for the conceptualizer, the reference point tends to be definite and discourse- or hearer-old information. This reference point ability has numerous linguistic manifestations, as shown in (9).

The possessor in (9a) and that book in (9b)–(c) must be familiar to the hearer to be felicitous; therefore, they are associated with hearer-old information. It is worth noting that while the topic is a reference point in the sense that it is a familiar entity to the interlocutor, not all reference points are topics. For example, the possessor within the NP in (9a) serves as the reference point with respect to the head nominal, but it cannot be a topic.

The correlation between reference point and topic becomes clearer when we encounter Korean examples, as in (10). Though grammatical encodings differ, all left-most nominals in (10) exhibit topicality in the sense that Cheli is given referentially as well as relationally. The sentences are about Cheli, and he is familiar to the hearer. Here, we denote topical entities in bold.

We argue that the overarching property of Cheli in (10a)–(c) is that it is a reference point; it becomes a prime candidate for a topic. The examples in (10) also demonstrate that different morphological markings can be used to denote a topic. While Chelswu is marked with different affixes in (10a) and (10b), it is construed as a topic in both examples.

As indicated, being a topic is sufficient for being a reference point; likewise, being a reference point is necessary for topicality. Let us now consider a more complicated example in (11). The answer, ku chayk-ul, is scrambled to the sentence-initial position. Here, the enclosed portion within a pair of square brackets denotes the focus, and the capital letters indicate that the entity exhibits prosodic prominence.

As ku chayk is the answer to the posed question in (11), it cannot be a topic.99.As we discuss in Section 3.1, ku chayk is an answer focus. However, ku chayk cannot be completely new to the hearer either; to make the answer felicitous, ku chayk must be hearer-old. Though non-topical, ku chayk gains a certain degree of prominence by appearing at the beginning of the sentence, which is generally a property manifested by a topical reference point. The question then is how ku chayk, as a non-topical element, is associated with the base clause informationally, and what motivates the scrambling of ku chayk. This is a puzzling problem the focalization in (11) presents. We discuss this issue in Section 4.

5.1The backmasking strategy

Backmasking refers to a technique in which a sound or message is recorded backward onto a track, which is meant to be played forward.1111.Wikipedia definition (accessed Feb 7, 2022). We refer to the phenomenon described in (19) as backmasking because the topical nominal Chelswu-nun/ka appears at the end of the sentence. As a topic, Chelswu-nun/ka is meant to be “about” the target clause.1212.Non-topical elements may be postposed, too. We discuss this issue in Section 6.

There is no denying that Chelswu is the topic in (19); the statement is about Chelswu, which invokes knowledge already in possession of the audience, and the statement is assessed as putative information about Chelswu. In addition, Chelswu is given both referentially and relationally. Nonetheless, Chelswu appears after the focus apeci-ka pwuca-ya ‘(someone’s) father is rich,’ which is non-canonical in Korean and in many other languages.

The utterance made in (19) is truth-conditionally identical to (20), where Chelswu gives rise to a topic. In (20), Chelswu serves as the reference point with respect to the inner clause apeci-ka pwuca-ya.

The reference point status of the first nominal in (20) has been extensively discussed in the literature, and interested readers should refer to the representative research, such as Kumashiro and Langacker (2003)Kumashiro, Toshiyuki, and Ronald W. Langacker 2003 “Double-Subject and Complex Predicate Constructions.” Cognitive Linguistics 14: 1–45. and Kumashiro (2016)Kumashiro, Toshiyuki 2016 A Cognitive Grammar of Japanese Clause Structure. Amsterdam: Benjamins. . These authors assume that the reference point appears in the left-most position, leaving utterances like (19) unexplained. While there is a way to resolve this puzzle using a technical apparatus within the framework these researchers adopt, we propose a simpler solution based on IS.

In Korean, a sentence does not require a grammatically encoded topic, particularly when the identification of a topic becomes possible for the addressee through a context. The first portion of the utterance in (19) – before the postposed topic – exemplifies such a case. Notwithstanding, the speaker may still provide a topic after the introduction of a focus to denote the mental address for the focus element. Providing an explicit topic is preferable to inferring it from a context because “guesswork” may not always be successful. For example, the topic of (19) without the postposed topic may not be clear to some people if the speaker utters the sentence looking at someone in the room; the hearer might interpret the sentence as a new episode (discourse) and construe the person the speaker looks at as the topic of this new episode. By providing the topic at the end of the utterance, the hearer is given an opportunity to “correct” her topic identification if it was not successful in her first attempt. If so, the backmasking strategy is an important tactic to make the hearer stay on the same page as the speaker. Note that although Chelswu appears at the end of the sentence in (19), its reference point function is maintained, and it is still the topic of the sentence; the IS of the utterance in (19) is simply encoded backward.

5.2The right dislocation strategy

Most researchers use the term right dislocation to include the phenomenon described in Section 5.1. However, we limit the term to the phenomenon where a resumptive pronoun occurs in the base clause, as in (21).

There exists a sizable amount of research on the right dislocation construction in Korean. The major concerns of the researchers revolve around how to syntactically associate the (null) pronoun with the extracted nominal, or how to identify the different types of right dislocation constructions, such as gapped vs. gapless (Ko 2014Ko, Heejeong 2014 Edges in Syntax: Scrambling and Cyclic Linearization. Oxford: Oxford University Press. , 2015Ko, Heejeong 2015 “Two Ways to the Right: A Hybrid Approach to Right Dislocation in Korean.” Language Research 51 (1): 3-40., 2016Ko, Heejeong 2016 “Gapless Right-Dislocation: The Role of Overt Correlates.” Language Research 52 (1): 3-32.; Ahn and Cho 2016Ahn, Hee-Don, and Sungeun Cho 2016 “A Uniform Analysis of Right-Dislocation: A Reply to Ko (2016).” Language Research 52 (5): 213–245; Furuya 2018Furuya, Kaori 2018 “A Reconsideration of the (Non-)uniform Syntax of Korean Right-Dislocation Description.” Linguistic Research 35(2): 275–304. among others). We set aside these issues for two reasons. First, the theoretical framework we assume – Cognitive Grammar – does not allow null elements or movements. Second, our main goal is to explicate how the form reflects its function.

We observe that a topical nominal tends to resist the right dislocation operation when the postposed entity is a bare nominal; while (22a)–(b) are awkward, (22c) is fully acceptable. By contrast, non-topical items, such as foci and adverbials, are much more susceptible to right dislocation, as illustrated in (22d)–(e).

In (22a), the pronominal ku pwun ‘that person’ is an irrefutable topic and therefore functions as a reference point with respect to the target proposition. It should be stressed that the conceptual content of the reference point, ku pwun, is only partially saturated in the base clause because pronominals are characteristically used deictically or anaphorically. The bare nominal, Kim-paksa-nim ‘Dr. Kim,’ may be construed as the reference point by establishing a correspondence relationship with the pronominal, sanctioning the fully saturated conceptual content of the reference point. Put differently, this correspondence relationship fulfills the anaphoric function of the pronominal. If the pronominal is used deictically, the postposed nominal may be construed as an add-on item, which elaborates the conceptual content of apeci ‘father’ in an appositive manner.1313.The add-on strategy is introduced in Section 6. These two competing strategies make the conceptualization of (22a) difficult for the hearer due to the failure of topic identification. A similar pattern is observed when the object is inanimate, as in (22b). Naomi may correspond to the established pronominal reference point or it may be associated with the focal entity, chayk ‘book’; e.g., Naomi may refer to Tanizaki’s famous novel. The rescue mechanism of (22a) is straightforwardly explained. When the postposed nominal is marked with -nun or -i, it is unambiguously associated with the pronominal, sanctioning a fully saturated conceptual content of the reference point. When marked with -nun, the postposed nominal in (22c) serves as a reference point for the target proposition; this is possible due to the same marker the topical and the postposed nominals adopt, as well as the topical pronominal’s partial conceptual content property. Note that (22a) may also be rescued with the nominative-marked postposed nominal because -i may also mark a topic, as we discussed earlier. The same rescue mechanism can be applied to (22b).1414.(22b) contrasts with the backmasking example shown in (19). In (19), as a relational noun, apeci ‘father’ invokes a reference point that in turn corresponds to a topical reference point for the target clause. Since the putative possessive nominal of apeci is not explicitly coded in the target clause, the postposed nominal naturally takes on that role. This is the only available construal for the postposed nominal; hence, the failure of topic identification does not arise even without the help of the topic marker.

Example (22d) tells a different story. The conceptual content of the reference point nominal, Chelswu, is fully saturated; therefore, the postposed nominal is not permitted to correspond to the reference point. In this example, the postposed nominal does not exhibit a direct association with the topic, and interpreting it as an add-on strategy does not become a viable option. Even if an association is possible between the topic and the postposed nominal, the postposed nominal cannot serve as a topical reference point because it does not access the main proposition. The naturalness of (22d) arises from the focal property of kukes ‘that thing’; as a focus, it presupposes an alternative set. Tellingly, the set of all individuals that can substitute for kukes must be presupposed in (22d). Due to the nature of an alternative set, a postposed focus performs an identification function; the postposed nominal in (22d) helps the addressee successfully identify kukes among the members in the presupposed set. The function of the postposition shown in (22e) is similar to that of (22d). Here, the exact manner is identified by the postposed adverbial. Example (22e) exhibits similar properties to the add-on strategy, which is discussed in the next section.

7.1Issues with FQs

Numerical classifiers in Korean occur in at least three different environments, as described in (25).

The floated quantifier type in which we are interested has been richly examined from a variety of theoretical perspectives. Broadly speaking, scholars are divided into two camps. The first view, often dubbed the stranding view, attempts to capture the similarities among the three types by deriving the FQ type from either the NI or the GC type. This view is supported by Miyagawa (1989)Miyagawa, Shigeru 1989 Structure and Case Marking in Japanese. NY: Academic Press. , Park and Sohn (1993)Park, Myung-Kwan, and Keun-Won Sohn 1993 “Floating Quantifiers, Scrambling, and the ECP.” In Japanese/Korean Linguistics, vol 3. ed. by Soonja Choi, 187–203. Stanford, CA: CSLI Publications., Choi (2001)Choi, Kiyong 2001 “hankwuke swulyangsa kwuseng-uy kwuco-wa uymi: pisokkyekhyeng-ul cwungsim-ulo [The Structure and Interpretation of Non-genitive Numeral Classifier Constructions in Korean].” Language Research 37 (3): 445–480., Kim (2005)Kim, Christina 2005 “Order and Meaning: Numeral Classifiers and Specificity in Korean.” In Proceedings of the 24th West Coast Conference on Formal Linguistics, ed. by John Alderete, 218–226. Somerville, MA: Cascadilla Proceedings Project., Ko (2007)Ko, Heejeong 2007 “Asymmetries in Scrambling and Cyclic Linearization.” Linguistic Inquiry 38: 49–83. , and Miyagawa and Arikawa (2007)Miyagawa, Shigeru, and Koji Arikawa 2007 “Locality in Syntax and Floated Quantifiers.” Linguistic Inquiry 38 (4): 645–670. among others. The second view, known as the VP-modifier view, does not assume this type of movement. Rather, the numeral classifier directly combines with a verbal predicate in syntax in the form of a head-modifier structure and semantically modifies the event structure of the predicate. This view is supported by Gunji and Hasida (1989)Gunji, Takao, and Koiti Hasida 1989 “Measurement and Quantification.” In Topics in Constraint-Based Grammar of Japanese, ed. by Takao Gunji, and Kôiti Hashida, 39–79. Dordrecth: Kluwer Academic Publishers., Fukushima (1991)Fukushima, Kazuhiko 1991 “Phrase Structure Grammar, Montague Semantics, and Floating Quantifiers in Japanese.” Linguistic and Philosophy 14: 581–628. , Kang (2002)Kang, Beom-Mo 2002 “Categories and Meanings of Korean Floating Quantifiers: With Some Reference to Japanese.” Journal of East Asian Linguistics 11: 375–398. , and Kim and Yang (2007)Kim, Jong-Bok, and Jaehyung Yang 2007 “Syntax and Semantics of Korean Numeral Classifier Constructions.” In Proceedings of the 14th International Conference on Head-Driven Phrase Structure Grammar, ed. by Stefan Müller, 136–172. Stanford: CSLI Online Publications. among others.

Though the evaluation of each approach is beyond the scope of this article, we would like to briefly discuss some weaknesses of the existing proposals, citing Kim (2013)Kim, Jong-Bok 2013 “Floated Numeral Classifiers in Korean: A Non-derivational, Functional Account.” Lingua 133: 189–212. . He provides an accurate assessment, as quoted below:

However, when we consider more data, one thing is clear that syntax alone is not enough to capture wider distributional possibilities of the FQ as well as speakers’ variations in the judgments of FQ data. The most serious challenge to both of these syntax-based views is the question of why the FQ “floats.”(Kim 2013Kim, Jong-Bok 2013 “Floated Numeral Classifiers in Korean: A Non-derivational, Functional Account.” Lingua 133: 189–212. , 201)

To overcome this challenge, Kim (2013)Kim, Jong-Bok 2013 “Floated Numeral Classifiers in Korean: A Non-derivational, Functional Account.” Lingua 133: 189–212. puts forward a third type of approach, which has a pragmatic orientation with an emphasis on IS. He argues that the floated quantifier functions as a focus marker and signals the partitioning of the thematic structure of the given sentence into theme and rheme. Kim (2013)Kim, Jong-Bok 2013 “Floated Numeral Classifiers in Korean: A Non-derivational, Functional Account.” Lingua 133: 189–212. uses these terms in the sense of Halliday and Matthiessen (2004)Halliday, Michael A. K., and Christian M. I. M. Matthiessen 2004 An Introduction to Functional Grammar. London, UK: Hodder Arnold.. For these scholars, the theme is the starting point of the message chosen by the speaker/writer, while the rheme is the remaining part that develops the theme, a participant, circumstance, or process. Note that the theme-rheme division does not always go hand-in-hand with the topic-comment division, though there are some overarching similarities. The gist of Kim’s proposal is given in (26), which can be rephrased as: quantifiers float to set off the rheme in the thematic structure.1616.For criticism of Kim (2013)Kim, Jong-Bok 2013 “Floated Numeral Classifiers in Korean: A Non-derivational, Functional Account.” Lingua 133: 189–212. , see Park and Yeon (2022)Park, Chongwon, and Jaehoon Yeon 2022 “Why Quantifiers Float: A Response to Kim (2013).” Lingua 277: 103403. .

Kim (2013)Kim, Jong-Bok 2013 “Floated Numeral Classifiers in Korean: A Non-derivational, Functional Account.” Lingua 133: 189–212. is a rare attempt to provide a functionalistic analysis of FQs in Korean. But since we are exploring this issue without relying on additional notions, such as theme and rheme, we need to find a different type of solution for the posed challenge. More importantly, Kim (2013)Kim, Jong-Bok 2013 “Floated Numeral Classifiers in Korean: A Non-derivational, Functional Account.” Lingua 133: 189–212. does not explore the relationship between the notion of focus and the givenness-related concepts, such as “in-focus” and “activation,” which we adopt in our analysis in the next section.

7.2Why do quantifiers float?

Let us first answer the question of why quantifiers float. We argue that quantifiers float to elevate an element currently not in focus to the in-focus state. One piece of evidence for our claim comes from examples in (27). While (27a) is not fully felicitous, the variations provided in (27b) and (27e) show significantly improved acceptability. In (27b), the FQ is accompanied by the focus particle, -man ‘only.’ Example (27c) illustrates a case of intervention effect that FQs induce. The FQ in (27c) cannot intervene between the Negative Polarity Item (NPI), ku chayk-pakkey ‘that book-only’ and its licensor, ahn-ass-ta ‘neg-pst-dcl.’ The same intervention effect is observed with a wh-expression which leads to an answer focus, as shown in (27d). The last example, (27e), demonstrates that (27a) may be rescued by placing prosodic prominence on the FQ, which is commonly observed in the realization of a focal entity in Korean. The examples in (27) therefore strongly signal that the FQ phenomenon is associated with the focus status of FQs; sentences with FQs are much more natural when the FQs exhibit focal properties in one way or another.

By severing the quantifier from its host, the quantifier is put in focus in two ways. First, it appears in a typical focus position. Second, it is associated with a typical focus element: the object.

Be that as it may, it is well-known that FQs may have subjects as their hosts, as in (28).

Shimojo (2004)Shimojo, Mitsuaki 2004 “Quantifier Float and Information Processing: A Case Study from Japanese.” Journal of Pragmatics 36: 375–405. accounts for the acceptability of (28) with (29), which is defined in the sub-conditions (a) and (b).

According to Shimojo, the quantifier seys ‘three’ in (28) can be scrambled only in the pre-object position. If it floats to the post-object position, the intervening element – the object – becomes not only an eligible host but also the preferred one.

Shimojo’s principles predict (30) will be unacceptable, and the prediction is indeed borne out.

It is important to note that Shimojo’s principles are applicable only to examples with bare FQs with neither a classifier nor a case marker. In (30), the quantifier is neutral with respect to animacy, and Shimojo’s principles work flawlessly because they are not sensitive to the markers that FQs carry.

Now let us consider B’s response in (31), which is a slightly revised version of (30); the person-denoting classifier is attached to the floated quantifier with prosodic prominence in (31). We believe the acceptability of B’s response improves under the context provided in (31), although the acceptability might be marginal for some speakers. While both haksayng and sey-myeng are brought into focus in B’s response, these constitute previously inactive information. As new information, they require specific cognitive effort to bring them into an activated stage. That is, the activation cost of haksayng is greater than that of maykcwu; therefore, the natural choice for the host of the quantifier becomes the subject nominal. With the prosodic prominence given to the quantifier in conjunction with its pre-verbal placement, the quantifier gives rise to a primary focus in B’s response in (31).

With (31), we have demonstrated that B’s response should not be judged out of context. In addition, the types of classifiers may affect the judgment regarding FQs.

As pointed out earlier, FQs in Korean may carry case markers. Let us consider (32), which is slightly different from B’s response in (31) in that the floated quantifier is marked nominative. Example (32) is fully acceptable with little contextual information because the nominative-marker of the floated quantifier strongly indicates its association with the subject nominal.

In the Introduction, we discussed Ko’s (2007) example identical to B’s response in (31): Ko judges B’s response as not felicitous at all.1717. Ko (2007)Ko, Heejeong 2007 “Asymmetries in Scrambling and Cyclic Linearization.” Linguistic Inquiry 38: 49–83. indicates that B’s response is not acceptable when the FQ is out of focus in (31). However, she does not discuss the rescue method we discuss here; it can be rescued by placing prosodic prominence on the FQ. While we agree with Ko (2007Ko, Heejeong 2007 “Asymmetries in Scrambling and Cyclic Linearization.” Linguistic Inquiry 38: 49–83. , 2018Ko, Heejeong 2018 “Scrambling in Korean Syntax.” Oxford Research Encyclopedias, 1–43 [online version]. ) that the degree of acceptability of B’s response in (31) is controversial, there is no denying that (33a–b) are either fully acceptable or exhibit a much higher degree of acceptability than B’s response in (31). Ko (2014)Ko, Heejeong 2014 Edges in Syntax: Scrambling and Cyclic Linearization. Oxford: Oxford University Press. and Ahn and Ko (2022)Ahn, Dorothy, and Heejeong Ko 2022 “On Non-conservativity of Korean Floating Quantifiers.” Glossa: A Journal of General Linguistics 7 (1): 1–34. deal with examples like (33) with syntactic mechanisms, which we do not endorse in our analysis.1818.In their analyses, the FQs in (33) are base-generated in a different position than that of the FQ in (27a).

The (improved) acceptability of (33a)–(b) naturally falls out in our analysis. The FQs in these examples are clearly marked with focus particles. As focus elements, they are previously inactive information, but they are explicitly brought into focus in these examples. With the help of the person-denoting classifier, the association between the quantifier and the subject nominal is established, where the subject nominal is an entity that exhibits a higher cost of activation.

One welcome outcome of our analysis concerns the (un)acceptability of the examples, as shown in (34a)–(b).

Examples (34a)–(b) illustrate FQs with different types of adverbs. As seen in (34a), a locative adverb may intervene between the floated quantifier and its host. Kim (2013)Kim, Jong-Bok 2013 “Floated Numeral Classifiers in Korean: A Non-derivational, Functional Account.” Lingua 133: 189–212. states that (34b), where a manner adverb appears between them, is marginally acceptable at best. Kuno and Takami (2003)Kuno, Susumu, and Ken-ichi Takami 2003 “Remarks on Uaccusativity and Unergativity in Japanese and Korean.” In Japanese/Korean Linguistics, vol. 12, ed. by William McClure, 280–294. Stanford, CA: CSLI Publications. provide a piece of supporting evidence for (34b) in reporting that manner adverbs are preferable as a focus. Since the manner adverb prefers to be a focus, it tends to have a higher cost of activation; then, (34b) becomes undesirable. It is worth mentioning that (34b) may be acceptable when it gives rise to a conservative reading with contrastiveness: (among five kids) three kids laughed loudly (but the other two didn’t).1919.For detailed discussion on (non-)conservative readings of FQs, please refer to Ahn and Sauerland (2017)Ahn, Dorothy, and Uli Sauerland 2017 “Measure Constructions with Relative Measures: Towards a Syntax of Non-conservative Construals.” The Linguistic Review 34 (2): 215-248. and Ahn and Ko (2022)Ahn, Dorothy, and Heejeong Ko 2022 “On Non-conservativity of Korean Floating Quantifiers.” Glossa: A Journal of General Linguistics 7 (1): 1–34. . Due to the limited space, we will leave the discussion on that reading for future research.