This dissertation explores, analyzes, and compares the usage of German and Chinese demonstratives. Discourse and textual uses of the forms will be considered as well as their locative and temporal uses. I observe that in both languages the demonstratives can be used to refer to referents. However, they depart from the common assumption that proximal demonstratives refer to entities or places close to the speaker and non-proximal demonstratives to entities or places far from the speaker. Having analyzed a language sample consisting of a German text and a Chinese text, I argue that both German and Chinese proximal demonstratives can signal the meaning of HIGH DEIXIS in a semantic system of DEIXIS in the Columbia School of Linguistics framework, whereas their non-proximal demonstratives can signal the meaning of LOW DEIXIS. In addition, Chinese demonstratives can be used under more circumstances than German demonstratives due to the lack of articles in Chinese. I also argue that Cognitive Linguistic analysis will better help a new language learner, whereas Columbia School Linguistics may be of greater assistance if the learner has advanced to a level where he/she needs to know more about the intrinsic differences between words with similar meanings and usages.

The German discourse marker na has mostly been neglected by particle researchers thus far. In German colloquial speech, it is frequently used, both as a stand-alone turn-initial marker ("Na, wenn du meinst" "Na, if you think so") and as part of several collocations such as na gut, na also, na dann, and na los. This dissertation uses a Conversation Analytic approach to find the function(s) and meaning(s) of the marker. We find that even though well is frequently provided as a translation for na, the two words do not share many functions.

Na-prefacing is used to qualify a statement and to return to a topic after going off on a tangent. Na can precede terms of endearment and conversation openers, and it precedes questions to make them sound more casual. Many of those na-prefaced turns are sequential departures where na alerts the hearer that something unexpected is coming. The collocations with na mostly have sequence management functions (sequential departure, sequence closure, topic shift).

Overall, na is a shift marker: It indicates that some kind of shift is coming up (a shift towards a new topic, a departure from the agenda of the conversation, etc.). The second component of the na-collocations allows for contextual fine-tuning, specifying the shift, and leading to a wider scope than the shifts indicated by a stand-alone na.

Pure synonymy is a myth. This is true in the lexicon of a language as well as its grammar. As such, why do grammars of Dutch label the following two forms below as synonymous?

Unsplit: Daarmee heb ik een probleem.

Therewith have I a problem.

Split: Daar heb ik een probleem mee.

There have I a problem with.

"I have a problem with that."

Calling these two forms above "the same" is an oversimplification. In my dissertation, Itest for difference in meaning between the forms and the motivation to choose one form over the other. To accomplish this goal, Iuse a corpus-based approach integrated with tests from Construction Grammar. We analyze these data in R (a statistics computing language) and Excel. This dissertation includes four tests, all complete. Preliminary findings underline a difference in meaning and provide motivation for the two forms. This is significant, as the current state of research suggests no real difference or meaning. Only one study, Honselaar (1994), discusses any possible variation in meaning. Additionally, emerging studies in linguistics frequently utilize larger data sets. This dissertation supports the shift toward more quantifiable data coupled with trained qualitative analysis. The quantifiable tests show a meaning difference, questioning the traditional analysis. Unsplit forms exhibit a shift in discourse, and split show continuation thereof.

Regarding motivation for the two forms, the data show clauses with verbs of low use frequency tend to be unsplit, suggesting that higher frequency verbs are more likely to occur with split forms. Frequency tends to be understood and decoded quicker (Fenk-Oczlon, 2001), and thus the listener/reader is generally able to parse the second half of the pronominal adverb, namely the preposition, before its utterance. This suggests that it would be easier to continue forward in the discourse if the form is split. These findings underline the growing need for further development of corpus-based linguistics studies. They welcome the movement of linguistics toward more empirical approaches to language.

This dissertation critically compares the two German focus particles eben and gerade. It has been repeatedly noted in the relevant literature that the two display an intriguing yet challenging near-synonymy. However, factors motivating this relationship have not been sufficiently explained to date. This study argues that the particles’ ostensible partial overlap is systematic and non-trivial in nature and that it can be explained by positing two distinct speaker motivations for uttering each particle to mark a constituent in a sentence: While the particle eben marks a constituent as conform-to-expectation, gerade marks a constituent as counter-to-expectation. Each marking is prompted by the discourse situation: If there is (extra)linguistic evidence that the interlocutor is inclined to select the same constituent as the speaker for completing a sentence, then eben is the appropriate marker. Alternatively, if no contrary evidence exists, eben may still be the appropriate marker. In the absense of favorable evidence, the speaker relies on assumptions about human communicative strategies and biases for arriving at the constituent choice matching the speaker’s and may still be prompted to utter eben. Conversely, the speaker utters gerade to mark a constituent as counter-to-expectation when there is (extra)linguistic evidence that a different constituent is more likely to be selected, or has been selected. Alternatively, if at least no (extra)linguistic evidence exists that the identical constituent is selected, assumptions about human communicative strategies and biases for arriving at a different constituent choice as the speaker’s prompt the use of gerade. The hypothesis is supported by two analyses: a minimal pair analysis employing substitution, elimination, and continuation tests as well as a synchronic, large-scale corpus-based analysis of syntactico-semantic features that frequently occur as the particles’ constituents. The former is employed to do away with compartmentalized interpretation suggestions for the particles, contaminated by adjacent elements, and suggests the underlying meanings of the particles to interact with said elements in generating the interpretations. The latter analysis presents a selection of syntactico-semantic features with which each particle frequently occurs in natural language and offers new insights on their functions based on interactions with eben and gerade.