Experimental investigations into the semantics of distributive marking
Bosnić, Ana
DOI:
10.33612/diss.171644158
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2021
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Bosnić, A. (2021). Experimental investigations into the semantics of distributive marking: Data from Serbian, Korean and Dutch. University of Groningen. https://doi.org/10.33612/diss.171644158
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
1
21
Our understanding of distributive markers across languages, although researched thoroughly, is still far from complete. Cross-linguistically, the ways in which distributive readings can be conveyed vary and are a fertile ground for comparative theoretical and experimental research. For example, some languages have distributive markers that appear to be very similar to universal distributive quantifiers such as every and
each, but turn out to have additional distributive interpretations, often
called event-distributive interpretations. In addition, they have different morpho-syntactic properties than universal distributive quantifiers – they mark the distributive share in a sentence, and are often referred to as
distributive share markers.
Broadly speaking, on the common assumption that the distributive markers in question are akin to universal distributive quantifiers, they would be analyzed as creating a relationship between two arguments in a sentence – the so-called distributive key (DistKey) argument and the
distributive share (DistShare) argument.1 While there is an abundance of experimental evidence on universal distributive quantifiers (also called
distributive key (DistKey) markers), especially with children (but also
with adults), experimental studies of DistShare markers remain scarce. This thesis contributes to the much needed experimental data on the adult interpretation of DistShare markers in Serbian2 and Korean, as well as acquisition data comparing languages with (Serbian) and without (Dutch) DistShare markers. The studies conducted here thus complement different theoretical accounts on DistShare markers developed over the years and highlight the importance of doing experimental investigations on the different semantic properties of these markers.
The DistKey/DistShare terminology is (partially) adopted from Choe (1987) and Gil (1995), who explicitly draw a parallel to universal quan-tifiers. In a nutshell, the DistKey argument is the argument that is being distributed over (similar to the restrictor of a quantifier), while the Dist-Share argument is the argument that is being distributed (similar to the
nuclear scope of a quantifier). Distributive share markers are syntactically
attached to the DistShare argument, hence the name. This property makes 1 This can be seen as a very simplified distinction between a distributional relationship. See especially Tovena (2016) and Cabredo-Hofherr & Tovena (2015) for a more refined terminology of key, sum-share,
k-units and sh-units. The distributive dependency may be established between whole sums (key and
sum-share) or individual units (k-units and sh-units). The key is the domain over which distribution takes place, and the k-unit is the singular or plural unit of the key that enters into the distributive relation (e.g., in (1) the key is the domain of children, and the k-unit is an individual child that will interact with the object). The sum-share is the plural expression that introduces the domain out of which sh-units are taken, meaning that the sh-unit is the singular or plural entity that gets paired with k-units (e.g., in (1), the sum-share is the domain of presents, and the sh-unit is an individual present that will be held by an individual child).
2 The reader should note that the Serbian language in linguistic research is also called Serbo-Croatian, Bosnian-Croatian-Serbian (BCS), or Bosnian-Croatian-Montenegrin-Serbian (BCMS), to indicate the four mutually intelligible standard varieties spoken in these countries. Since I am the speaker of the Serbian variety and I did not test speakers of other dialects to confirm my results, I simply refer to the language as Serbian (although I strongly suspect the results to be transferrable).
1
1 Introduction
Our understanding of distributive markers across languages, although researched thoroughly, is still far from complete. Cross-linguistically, the ways in which distributive readings can be conveyed vary and are a fertile ground for comparative theoretical and experimental research. For example, some languages have distributive markers that appear to be very similar to universal distributive quantifiers such as every and
each, but turn out to have additional distributive interpretations, often
called event-distributive interpretations. In addition, they have different morpho-syntactic properties than universal distributive quantifiers – they mark the distributive share in a sentence, and are often referred to as
distributive share markers.
Broadly speaking, on the common assumption that the distributive markers in question are akin to universal distributive quantifiers, they would be analyzed as creating a relationship between two arguments in a sentence – the so-called distributive key (DistKey) argument and the
distributive share (DistShare) argument.1 While there is an abundance of experimental evidence on universal distributive quantifiers (also called
distributive key (DistKey) markers), especially with children (but also
with adults), experimental studies of DistShare markers remain scarce. This thesis contributes to the much needed experimental data on the adult
interpretation of DistShare markers in Serbian2 and Korean, as well as acquisition data comparing languages with (Serbian) and without (Dutch) DistShare markers. The studies conducted here thus complement different theoretical accounts on DistShare markers developed over the years and highlight the importance of doing experimental investigations on the different semantic properties of these markers.
The DistKey/DistShare terminology is (partially) adopted from Choe (1987) and Gil (1995), who explicitly draw a parallel to universal quan-tifiers. In a nutshell, the DistKey argument is the argument that is being distributed over (similar to the restrictor of a quantifier), while the Dist-Share argument is the argument that is being distributed (similar to the
nuclear scope of a quantifier). Distributive share markers are syntactically
attached to the DistShare argument, hence the name. This property makes 1 This can be seen as a very simplified distinction between a distributional relationship. See especially Tovena (2016) and Cabredo-Hofherr & Tovena (2015) for a more refined terminology of key, sum-share,
k-units and sh-units. The distributive dependency may be established between whole sums (key and
sum-share) or individual units (k-units and sh-units). The key is the domain over which distribution takes place, and the k-unit is the singular or plural unit of the key that enters into the distributive relation (e.g., in (1) the key is the domain of children, and the k-unit is an individual child that will interact with the object). The sum-share is the plural expression that introduces the domain out of which sh-units are taken, meaning that the sh-unit is the singular or plural entity that gets paired with k-units (e.g., in (1), the sum-share is the domain of presents, and the sh-unit is an individual present that will be held by an individual child).
2 The reader should note that the Serbian language in linguistic research is also called Serbo-Croatian, Bosnian-Croatian-Serbian (BCS), or Bosnian-Croatian-Montenegrin-Serbian (BCMS), to indicate the four mutually intelligible standard varieties spoken in these countries. Since I am the speaker of the Serbian variety and I did not test speakers of other dialects to confirm my results, I simply refer to the language as Serbian (although I strongly suspect the results to be transferrable).
1
23 22
DistKey argument. Thus, this particular terminology serves as the most straight-forward typological and morpho-syntactic distinction between two major types of distributive markers.
Let us first illustrate this phenomenon with an example from English and Serbian in (1):
(1) a. Each child is holding a present. – English b. Deca drže po jedan poklon. – Serbian
children.nom hold.pl distr one present.acc ‘The children are holding (distributively) one present.’
Here, the distributive quantifier each syntactically attaches to the ar-gument that is being distributed over (i.e., child). On the other hand, the distributive marker po in Serbian attaches to the argument that is being distributed (i.e., presents). Furthermore, there are semantic differences between these two sentences. Apart from a shared reading in which there is one present for each individual child (a.k.a. individual-distributive
read-ing), the example from Serbian also yields event-distributive readings, in
which the DistShare argument can be distributed over spatial locations or temporal units (e.g., there could be several groups of children in different places holding one present together).
Throughout this thesis, we exclusively use the term distributive share
(DistShare) markers, as it is the most appropriate term that specifically
high-lights the typological and morpho-syntactic properties of these markers we are interested in. By using this term, however, we do not mean to commit to an analysis that the DistShare markers semantically entail set-relational, quantificational relationship between the arguments (something we are investigating), but simply use it as a useful term that reflects the typological classification, to talk about the underlying semantics of these markers. 1.1 The scope of the thesis
The main focus of this thesis is on different interpretations of DistShare markers and experimentally teasing apart the predictions of opposite semantic analyses proposed for them. While these analyses agree that DistShare markers yield event-distributive readings, there is a crucial divergence between them, which essentially results into two lines of research – one approach is to analyze DistShare markers as universal quantifiers that can distribute over events, while the other approach has weaker requirements for these markers and analyzes them as distributive markers of event plurality (pluractionals) without universal quantification. This distinction is pivotal to defining the scope of this thesis. We dedicate a large portion of our research to experimentally check which approach is most appropriate for DistShare markers in two languages – Serbian and Korean. In addition, this thesis will also connect the adult studies to acquisition results, mapping the path of acquiring DistShare and DistKey markers in Serbian and Dutch.
has been tackled from many different directions and in different languages, and for this reason we actually have quite a diverse set of results and theories. This can also be seen from diverse terminology used to refer to DistShare markers: distributive numerals, anti-quantifiers, distance
distrib-utive markers or dependent indefinites. However, these terms carry specific
properties to these markers that often do not fully overlap, especially with the languages tested here. Below, we present a few reasons why we are not using the above-mentioned terms (and their accompanying analyses):
First, the term distributive numeral was first introduced by Gil (1982) in his dissertation on distributivity and different types of distributive markers. Distributive numerals are modified numeral phrases that force distributive readings. For instance, the expressions like Latin terni, Serbian po tri, Romanian câte trei, Georgian sam-sami, correspond to expressions that can be glossed as in threes, sets of three, three at a time or three each in English. The term was also used in Balusu’s work on Telugu (2006), Cable’s work on Tlingit (2014) and Cabredo-Hofherr & Etxeberria’s work on Basque (2017), among others. However, even though the term distributive numeral is the closest to the properties to the markers in Serbian and Korean, it is a narrower term, since a marker such as Serbian po can also attach to weak quantifiers, adjectives and adverbs.
Second, the term anti-quantifier (introduced by Choe (1987)) or
distance-distributivity (used by authors like Zimmermann (2002b) or
Champollion (2012; 2015; 2016b)) could be initially misleading for us. For instance, Zimmermann (2002b) claims that distance-distributive elements should be analyzed as universal distributive quantifiers, but that these elements occur at a distance from their restriction, a.k.a. the DistKey argument. This, however, is precisely the theoretical assumption we are experimentally testing in this thesis and discussing in detail in Chapter 3. The term thus encompasses markers such as English
binominal each (e.g., The children are holding one present each), which,
although syntactically associated with the distributive share one present, it is semantically associated with the DistKey argument the children at a distance. In addition, here we are not concerned with the analysis of binominal each and we are distinguishing this type of marker from markers such as Serbian po or Korean -ssik.
Finally, the term dependent indefinites is used by authors like Farkas (1997, 2015), Henderson (2011, 2014) and Kuhn (2015, 2017), talking about markers that associate with the distributive share argument. The crucial difference between dependent indefinites (and dependent numerals, as a subtype) and markers in Serbian and Korean is that dependent indefinites are analyzed as distributive operators that must have an overt plural DP licensor, or be in the scope of a quantifier (Farkas 1997, Henderson 2011). What follows from this analysis is that event-distributive readings achieved via linguistically implicit DistKey are not possible, which does not hold for Serbian po and Korean -ssik. For this reason, we will also not adopt this term in our work.
1
these markers (at least syntactically) different from those that attach to the DistKey argument. Thus, this particular terminology serves as the most straight-forward typological and morpho-syntactic distinction between two major types of distributive markers.
Let us first illustrate this phenomenon with an example from English and Serbian in (1):
(1) a. Each child is holding a present. – English b. Deca drže po jedan poklon. – Serbian
children.nom hold.pl distr one present.acc ‘The children are holding (distributively) one present.’
Here, the distributive quantifier each syntactically attaches to the ar-gument that is being distributed over (i.e., child). On the other hand, the distributive marker po in Serbian attaches to the argument that is being distributed (i.e., presents). Furthermore, there are semantic differences between these two sentences. Apart from a shared reading in which there is one present for each individual child (a.k.a. individual-distributive
read-ing), the example from Serbian also yields event-distributive readings, in
which the DistShare argument can be distributed over spatial locations or temporal units (e.g., there could be several groups of children in different places holding one present together).
Throughout this thesis, we exclusively use the term distributive share
(DistShare) markers, as it is the most appropriate term that specifically
high-lights the typological and morpho-syntactic properties of these markers we are interested in. By using this term, however, we do not mean to commit to an analysis that the DistShare markers semantically entail set-relational, quantificational relationship between the arguments (something we are investigating), but simply use it as a useful term that reflects the typological classification, to talk about the underlying semantics of these markers. 1.1 The scope of the thesis
The main focus of this thesis is on different interpretations of DistShare markers and experimentally teasing apart the predictions of opposite semantic analyses proposed for them. While these analyses agree that DistShare markers yield event-distributive readings, there is a crucial divergence between them, which essentially results into two lines of research – one approach is to analyze DistShare markers as universal quantifiers that can distribute over events, while the other approach has weaker requirements for these markers and analyzes them as distributive markers of event plurality (pluractionals) without universal quantification. This distinction is pivotal to defining the scope of this thesis. We dedicate a large portion of our research to experimentally check which approach is most appropriate for DistShare markers in two languages – Serbian and Korean. In addition, this thesis will also connect the adult studies to acquisition results, mapping the path of acquiring DistShare and DistKey markers in Serbian and Dutch.
However, it is important to mention that the topic of DistShare markers has been tackled from many different directions and in different languages, and for this reason we actually have quite a diverse set of results and theories. This can also be seen from diverse terminology used to refer to DistShare markers: distributive numerals, anti-quantifiers, distance
distrib-utive markers or dependent indefinites. However, these terms carry specific
properties to these markers that often do not fully overlap, especially with the languages tested here. Below, we present a few reasons why we are not using the above-mentioned terms (and their accompanying analyses):
First, the term distributive numeral was first introduced by Gil (1982) in his dissertation on distributivity and different types of distributive markers. Distributive numerals are modified numeral phrases that force distributive readings. For instance, the expressions like Latin terni, Serbian po tri, Romanian câte trei, Georgian sam-sami, correspond to expressions that can be glossed as in threes, sets of three, three at a time or three each in English. The term was also used in Balusu’s work on Telugu (2006), Cable’s work on Tlingit (2014) and Cabredo-Hofherr & Etxeberria’s work on Basque (2017), among others. However, even though the term distributive numeral is the closest to the properties to the markers in Serbian and Korean, it is a narrower term, since a marker such as Serbian po can also attach to weak quantifiers, adjectives and adverbs.
Second, the term anti-quantifier (introduced by Choe (1987)) or
distance-distributivity (used by authors like Zimmermann (2002b) or
Champollion (2012; 2015; 2016b)) could be initially misleading for us. For instance, Zimmermann (2002b) claims that distance-distributive elements should be analyzed as universal distributive quantifiers, but that these elements occur at a distance from their restriction, a.k.a. the DistKey argument. This, however, is precisely the theoretical assumption we are experimentally testing in this thesis and discussing in detail in Chapter 3. The term thus encompasses markers such as English
binominal each (e.g., The children are holding one present each), which,
although syntactically associated with the distributive share one present, it is semantically associated with the DistKey argument the children at a distance. In addition, here we are not concerned with the analysis of binominal each and we are distinguishing this type of marker from markers such as Serbian po or Korean -ssik.
Finally, the term dependent indefinites is used by authors like Farkas (1997, 2015), Henderson (2011, 2014) and Kuhn (2015, 2017), talking about markers that associate with the distributive share argument. The crucial difference between dependent indefinites (and dependent numerals, as a subtype) and markers in Serbian and Korean is that dependent indefinites are analyzed as distributive operators that must have an overt plural DP licensor, or be in the scope of a quantifier (Farkas 1997, Henderson 2011). What follows from this analysis is that event-distributive readings achieved via linguistically implicit DistKey are not possible, which does not hold for Serbian po and Korean -ssik. For this reason, we will also not adopt this term in our work.
1
25 24
This thesis is comprised of 6 chapters (including the Introduction and Con-clusion), three of which are in the form of independent research articles, each dealing with a specific research question. The chapters range from act-out tasks in a language development study and experimental studies developed to test adult interpretations of sentences with DistShare markers, to a discussion of the theoretical implications of our experimental results and establishing ground for further research. The book is organized as follows:
Chapter 2 presents data from a child language study done in Serbian and Dutch. The purpose of this study was to test how the individual-distributive readings differ between languages with and without DistShare markers. Using an act-out task, we elicited answers from children from ages 7-9 to determine their comprehension of distributively marked sentences and sentences with bare numerals, as well as their performance in understand-ing distributive readunderstand-ings. Apart from distributively unmarked sentences, for Serbian, we used sentences with either a distributive universal quan-tifier svaki, or with a DistShare marker po, while for Dutch we only used sentences with a distributive universal quantifier elke. The findings show that Serbian children acquire po considerably later than the quantifier
svaki – around the ages of 9 or 10. In addition, they also lag behind fully
understanding the universal quantifier svaki in comparison with their Dutch peers and their understanding of elke (which is already mastered around the age of 6). By doing an act-out task, we also caught a glimpse of children’s reasoning about distributivity and a possible learning trajectory going from preferring distributive interpretations for unmarked sentences to understanding distributive markers. Finally, we observed pragmatic differences between Dutch and Serbian response patterns – Dutch adults seem to be influenced by pragmatic reasons when giving their responses for numerically marked sentences and show a wider variety of possible responses, while Serbian adults are not influenced by pragmatic factors.
Chapter 3 moves on to adult interpretations of distributively marked sen-tences in Serbian and Korean. The experiments were designed to examine a core semantic issue: are DistShare markers true universal quantifiers, or are they event plurality markers? The latter means that they simply require plurality of events rather than additional universal properties of quantifiers, such as an exhausted DistKey. Both languages, although typologically very distant, have DistShare markers that share many semantic properties with each other. Using a picture verification task and intransitive sentences in which the only available argument has a DistShare marker attached, we wanted to test whether there was then a spatial DistKey present at all, since its presence is a defining property of universal quantifiers. If there was a spatial DistKey, it had to be implicit and, if DistShare markers are universal quantifiers, this implicit DistKey would have to be exhausted. The results revealed that these markers have universal quantificational force, establishing a distributive relation between the events and the implicit DistKey, but crucially not between overt atomic units. Importantly, we
atomic) individuals given in the visual input of the experimental conditions. We analyzed this as universal event-distributive quantification over spatial units that were defined by these groups of individuals. This then means that the spatial DistKey is divided into relevant parts that need to be exhausted by the DistShare argument.
Chapter 4 presents a three-part study designed to test the predictions put forth in Chapter 3 using transitive sentences with the DistShare marker
po attached to the object argument. Specifically, we wanted to examine
whether (and to what degree) the predictions hold in cases where there is an overt competitor for a DistKey (i.e., the subject argument). We con-ducted three experiments (picture verification tasks) and tested Serbian adult speakers to establish the baselines about individual-distributive and event-distributive readings. However, we unexpectedly found evidence of two populations of speakers – one that seems to interpret po as a universal quantifier (over events) and the other that seems to interpret it as a plu-ractional marker, which we refer to as “non-exhaustive po”, since it did not show exhaustivity requirements. We close this chapter by speculating that po may be undergoing a semantic change where, for some speakers,
po has lost its universal quantificational force.
Chapter 5 is a further theoretical analysis of the results we have found in Chapter 4, specifically. The focus of the chapter is exclusively on the interpretation of “non-exhaustive” po that does not require a DistKey to be exhausted, how to semantically analyze it and how it relates to the existing body of literature that goes beyond universal quantifiers. We selected two theoretical approaches to discuss as potential explanations for non-exhaus-tive po. The divergence between the approaches is whether non-exhausnon-exhaus-tive
po necessarily results in multiple (sum) of events (pluractional account)
or not (group-forming account). We then provide empirical evidence that distribution can be done over three dimensions – individuals, time and space. We introduce crucial empirical examples, which we also take to be the test cases for counting minimally different events. The discussion turns in favor for a pluractional account, the one that assumes multiple events as the only (weak) requirement for felicitous situations with po. The proposal for future research is to investigate further into the directionality of semantic change of po and, depending whether non-exhaustive po is the result of weakening and strengthening, develop a unified account of this and other similar markers. This could potentially be achieved by weakening the universal quantifier proposal or strengthening the pluractional one. Lastly, the aspect of differentiating the dimensions over which distribution can take place may be included in the semantics of these markers.
The thesis is concluded and summarized with Chapter 6 with an ad-ditional focus on further research and open questions that arose during different stages of the present work.
1
1.2 The roadmap
This thesis is comprised of 6 chapters (including the Introduction and Con-clusion), three of which are in the form of independent research articles, each dealing with a specific research question. The chapters range from act-out tasks in a language development study and experimental studies developed to test adult interpretations of sentences with DistShare markers, to a discussion of the theoretical implications of our experimental results and establishing ground for further research. The book is organized as follows:
Chapter 2 presents data from a child language study done in Serbian and Dutch. The purpose of this study was to test how the individual-distributive readings differ between languages with and without DistShare markers. Using an act-out task, we elicited answers from children from ages 7-9 to determine their comprehension of distributively marked sentences and sentences with bare numerals, as well as their performance in understand-ing distributive readunderstand-ings. Apart from distributively unmarked sentences, for Serbian, we used sentences with either a distributive universal quan-tifier svaki, or with a DistShare marker po, while for Dutch we only used sentences with a distributive universal quantifier elke. The findings show that Serbian children acquire po considerably later than the quantifier
svaki – around the ages of 9 or 10. In addition, they also lag behind fully
understanding the universal quantifier svaki in comparison with their Dutch peers and their understanding of elke (which is already mastered around the age of 6). By doing an act-out task, we also caught a glimpse of children’s reasoning about distributivity and a possible learning trajectory going from preferring distributive interpretations for unmarked sentences to understanding distributive markers. Finally, we observed pragmatic differences between Dutch and Serbian response patterns – Dutch adults seem to be influenced by pragmatic reasons when giving their responses for numerically marked sentences and show a wider variety of possible responses, while Serbian adults are not influenced by pragmatic factors.
Chapter 3 moves on to adult interpretations of distributively marked sen-tences in Serbian and Korean. The experiments were designed to examine a core semantic issue: are DistShare markers true universal quantifiers, or are they event plurality markers? The latter means that they simply require plurality of events rather than additional universal properties of quantifiers, such as an exhausted DistKey. Both languages, although typologically very distant, have DistShare markers that share many semantic properties with each other. Using a picture verification task and intransitive sentences in which the only available argument has a DistShare marker attached, we wanted to test whether there was then a spatial DistKey present at all, since its presence is a defining property of universal quantifiers. If there was a spatial DistKey, it had to be implicit and, if DistShare markers are universal quantifiers, this implicit DistKey would have to be exhausted. The results revealed that these markers have universal quantificational force, establishing a distributive relation between the events and the implicit DistKey, but crucially not between overt atomic units. Importantly, we
discovered the implicit DistKey had to be defined by the groups of (non-atomic) individuals given in the visual input of the experimental conditions. We analyzed this as universal event-distributive quantification over spatial units that were defined by these groups of individuals. This then means that the spatial DistKey is divided into relevant parts that need to be exhausted by the DistShare argument.
Chapter 4 presents a three-part study designed to test the predictions put forth in Chapter 3 using transitive sentences with the DistShare marker
po attached to the object argument. Specifically, we wanted to examine
whether (and to what degree) the predictions hold in cases where there is an overt competitor for a DistKey (i.e., the subject argument). We con-ducted three experiments (picture verification tasks) and tested Serbian adult speakers to establish the baselines about individual-distributive and event-distributive readings. However, we unexpectedly found evidence of two populations of speakers – one that seems to interpret po as a universal quantifier (over events) and the other that seems to interpret it as a plu-ractional marker, which we refer to as “non-exhaustive po”, since it did not show exhaustivity requirements. We close this chapter by speculating that po may be undergoing a semantic change where, for some speakers,
po has lost its universal quantificational force.
Chapter 5 is a further theoretical analysis of the results we have found in Chapter 4, specifically. The focus of the chapter is exclusively on the interpretation of “non-exhaustive” po that does not require a DistKey to be exhausted, how to semantically analyze it and how it relates to the existing body of literature that goes beyond universal quantifiers. We selected two theoretical approaches to discuss as potential explanations for non-exhaus-tive po. The divergence between the approaches is whether non-exhausnon-exhaus-tive
po necessarily results in multiple (sum) of events (pluractional account)
or not (group-forming account). We then provide empirical evidence that distribution can be done over three dimensions – individuals, time and space. We introduce crucial empirical examples, which we also take to be the test cases for counting minimally different events. The discussion turns in favor for a pluractional account, the one that assumes multiple events as the only (weak) requirement for felicitous situations with po. The proposal for future research is to investigate further into the directionality of semantic change of po and, depending whether non-exhaustive po is the result of weakening and strengthening, develop a unified account of this and other similar markers. This could potentially be achieved by weakening the universal quantifier proposal or strengthening the pluractional one. Lastly, the aspect of differentiating the dimensions over which distribution can take place may be included in the semantics of these markers.
The thesis is concluded and summarized with Chapter 6 with an ad-ditional focus on further research and open questions that arose during different stages of the present work.
in Serbian and Dutch – evidence from an
act-out task
This chapter was published in a shorter form as:
Bosnić, A. & Spenader, J. (2019). Acquisition path of distributive markers in Serbian and Dutch: Evidence from an act-out task. In Brown, M. M. & Dailey, B (eds.). BUCLD 43: Proceedings of the 43rd annual Boston University Conference on Language Development. Vol.1, 94-108. Cascadilla Press.
