- 6 - Jufrizal
- Cited by other publications
- Passivization And Typology: Form And Function (Typological Studies in Language)
- Passivization and Typology: Form and function | Edited by Werner Abraham and Larisa Leisiö
Each item was two sentences in length.
6 - Jufrizal
As before, the second sentence described a transitive event, while the first provided a preamble. The first manipulated factor, information status , was established as in Experiment 1. It was him who a thief attacked upon arriving. It was a thief who attacked him upon arriving. It was a cowboy who she attacked upon arriving.
It was her who attacked a cowboy upon arriving. In addition to this, the second sentence featured a cleft containing one of the protagonists; that is, either the Agent or Patient of the critical transitive event was foregrounded. This placed a clear focus on that protagonist and constituted the second manipulated factor; focus. The transitive description was always in active-voice; this meant that any changes in passive production would be motivated by the experimental factors.
The procedure followed that of Experiment 1. This resulted in six items per condition per file, ensuring an equal frequency of each condition in each file. Responses were coded for voice at three levels active, be-passive, or get-passive , and passive responses were also coded for by-phrase inclusion. C ross-tabulation results , giving raw counts of responses AV: active-voice, B0: be-passive without by-phrase, BB: be-passive with by-phrase, G0: get-passive without by-phrase, GB: get-passive with by-phrase for the two levels per factor Information Status and Focus.
Get-passives are the least frequent form overall, though there is a notable increase in their likelihood over Experiment 1, with most of these appearing in focussed Patient conditions. The inferential analyses below focused on a the general likelihood of producing a passive voice paraphrase, b the likelihood of producing a be-passive paraphrase, and c the likelihood of producing a get-passive paraphrase, respectively.
Cited by other publications
For this stage of analysis, all passive responses be-passives and get-passives with and without an agentive by-phrase were combined into a single category. Again, we used binary logistic GEEs by participants and items, assuming an exchangeable covariance structure for repeated measurements to model occurrences of passive-voice over active voice paraphrases as a function of Information Status and Focus , as well as the interaction between them.
There was a main effect of Information Status , such that passive-voice paraphrases were reliably more likely when the Patient was given 0. There was also a main effect of Focus significant by items only , whereby passive-voice responses were more likely when the Patient was focussed via clefting 0. Next, we considered the production of be-passive paraphrases out of all valid responses active-voice, be-passives, and get-passives.
The model was used to predict likelihood of producing a be-passive response as a function of Information Status and Focus , as well as the interaction between them. There was no effect of Focus. In this analysis we considered the production of get-passive paraphrases out of all valid responses.
The model was used to predict the likelihood of producing a get-passive response as a function of Information Status and Focus , as well as the interaction between them. This effect was marginal by items, and significant by subjects. There was no effect of Information Status.
As in the first experiment, Experiment 2 revealed fewer passive-voice responses overall than active-voice responses. The main effect of Information Status was also maintained from Experiment 1: a given Patient increased the likelihood of passive-voice paraphrases in complementary distribution with active-voice paraphrases, which were more likely following a given Agent. Importantly, analyses focussing on each passive-type separately indicated that this effect of Information Status mainly affected the likelihood of the more canonical be-passive form rather than the less common get-passive form as was also suggested in Experiment 1 where get-passive occurrences were very rare.
The additional focus manipulation via clefting of either Agent or Patient did not interact with Information Status, but independently influenced responses. Interestingly, in contrast to Information Status, the focus manipulation mainly affected occurrences of get-passive rather than be-passive paraphrases: regardless of which protagonist was given , get-passive production was more likely following a story with a clefted Patient in the last sentence rather than a clefted Agent.
Given the low probability of get-passive uses in general, it remains to be seen how well this latter finding would replicate in future research. However, it does lend some support to theories which claim that emphasis on the Patient is a significant contributor to the production of get-passives. By contrast, probabilities of be-passives appear to be much less effected by variations in Patient-related focus.
Experiments 1 and 2 both indicated that, relative to a given Agent, a given Patient reliably promotes the use of passive-voice, and of be-passives in particular. In Experiment 2 we further showed that the use of get-passives, but not the use of be-passives, is promoted by a general focus or mark of importance on the Patient of an action via clefting. In Experiment 3 below, we consider the effect of a more specific Patient-related attribute. As noted earlier, existing literature offers multiple suggestions as to what constitutes the main factor promoting get-passive use; the majority of these suggestions, however, do gravitate toward the event's Patient.
Among these diverse proposals, there are several that fall within the attribution of agentivity of the Patient, including Patient responsibility or blame, Patient purposefulness, Patient initiative or control, etc. Here, we subsume the above into the category of agentivity , using Patient-related questions for paraphrasing. In those questions, we either frame the Patient as being more Patient-like a passive undergoer of an action; e. These Patient-related questions are also likely to elicit a greater number of passive-voice responses overall thus increasing the reliability of statistical interpretation.
These twenty-four participants had not participated in either of the previous experiments.soilstones.com/wp-content/2020-03-04/1270.php
Passivization And Typology: Form And Function (Typological Studies in Language)
Twenty-four sets of materials were created based on those used in Experiment 1. Here, each had four conditional variants as in 5. As in both previous experiments, the second sentence described a transitive event, while the first provided a preamble. The first manipulated factor, Information Status , was established as in Experiments 1 and 2.
In addition, rather than having a simple visual prompt to retell the event, participants were presented with one of two types of question, framing the Patient of the action either as being Patient-like What happened to the Patient?
Passivization and Typology: Form and function | Edited by Werner Abraham and Larisa Leisiö
This constituted the second manipulated factor; Patient Framing. The critical transitive description per story was always in active-voice, ensuring that any changes in passive-voice production would be motivated by the experimental factors. As stated throughout, we were concerned with the production of syntactic alternatives, and deemed it necessary for participants to demonstrate the availability of at least two syntactic forms.
In this instance, all participants produced at least two alternatives, and were therefore included in the subsequent analyses. Less than 0. C ross-tabulation results , giving raw counts of responses AV, active-voice; B0, be-passive without by-phrase; BB, be-passive with by-phrase; G0, get-passive without by-phrase; GB, get-passive with by-phrase for the two levels per factor Information Status and Patient Framing.
Active-voice responses are no longer dominant, and most actives now appear following a given Patient as opposed to a given Agent in the first two experiments. Be-passives now take over as by-far the most frequent response type; contrary to both Experiments 1 and 2, most be-passives appear following a given Agent , rather than a given Patient as previously seen. Get-passives are the least frequent form overall, though there is a notable increase in their absolute frequency over Experiments 1 and 2, with most get-passives appearing in given Patient conditions the conditions in which be-passives were previously observed to be more likely.
As before, data were analyzed using binary logistic GEEs by participants and items treating Information Status and Patient Framing as repeated-measures predictors with exchangeable covariance structure. In the first analysis, all passive responses be-passives and get-passives with and without a by-phrase were combined into a single category and the models were set up to predict occurrences of passive-voice paraphrases as a function of Information Status and Patient Framing , as well as the interaction between them.
There was also a main effect of Patient Framing , whereby passive-voice responses were more likely when the questions framed the Patient as more Patient-like 0. Next, we considered the production of be-passive paraphrases out of all valid responses, i. As indicated, there was a main effect of Information Status , again the inverse of that seen in the previous experiments: a given Agent reliably increased the likelihood of a be-passive paraphrase 0. There was also a main effect of Patient Framing , whereby be-passives were significantly more frequent after a question framing the Patient as Patient-like 0.
The interaction did not approach significance. In this analysis we considered the production of get-passives out of all valid responses. The GEE model was used to predict likelihood of producing a get-passive response as a function of Information Status and Patient Framing , as well as the interaction between them. As shown, there was a main effect of Information Status , which was absent from the data in Experiment 2: get-passive paraphrases were reliably more likely after a given Patient 0.
This effect was only marginal by subjects, but significant by items. There was no effect of Framing. In Experiments 1 and 2, we have not reported analyses of by-phrase inclusion, since the relevant descriptive statistics indicated that it was rather uncommon for participants to drop the by-phrase, and hence inferential statistics could not be informatively applied.
However, Experiment 3 elicited a far greater number of passives overall, allowing a more informative consideration of by-phrase inclusion. The binary logistic GEE models were used to predict the likelihood of including an agentive by-phrase in a passive response as a function of Information Status, Patient Framing , and Passive-type whether participants used be or get to form their passive response , and all possible interactions between these factors. Generalized-Score Chi-Squares from binary logistic GEE analyses, predicting the likelihood of agentive by-phrase inclusion by factor combinations of Information Status given Agent or given Patient , Patient Framing Agent-like or Patient-like , and Passive-type be or get.
As is evident, only one significant effect was established in this analysis, namely a main effect of Passive-type , whereby by-inclusion was reliably increased when the passive-type used was be 0. The main effect of Information Status that was observed in the first two experiments is entirely reversed in Experiment 3. That is, a given Agent now increased the likelihood of passive-voice rather than active-voice , while a given Patient increased the likelihood of active-voice responses.
Note that this effect is mainly driven by be-passive responses which now constitute the majority of responses overall , whereas the effect of Information Status on get-passives is smaller, and in the opposite direction, therefore comparable to the effect of Information Status on be-passives in Experiments 1 and 2. The latter is interesting since Information Status previously had no effect on get-passive likelihood, but now it displays a main effect such that get-passive paraphrases become more likely when the Patient is given comparable to the effect of Information Status previously observed for be-passive uses.
In other words, the effect of Information Status on be-passive uses is now reversed, yet the original effect appears to have transferred to get-passive uses, as if get is now filling the role previously filled by be. Clearly, the Patient-related questions in Experiment 3 as opposed to free paraphrasing in Experiments 1 and 2 must be at least partly responsible for these changes in the effects of Information Status. Consider that the Patient-related questions in Experiment 3 actually introduced additional information to the discourse established by the story.
- Passivization And Typology: Form And Function (Typological Studies in Language) - PDF Free Download.
- Patient-related constraints on get- and be-passive uses in English: evidence from paraphrasing.
- Nessus Network Auditing.
While the story establishes which of the two protagonists Agent or Patient is given vs. This in turn means that the Information Status manipulation in the story no longer acts as a cue for subject-assignment, and hence passive vs.
- brian nolan - Google Scholar Citations;
- Plant Hormones: Physiology, Biochemistry and Molecular Biology.
- Heidegger for Architects (Thinkers for Architects).
- Help Desk/Feedback.
Further evidence for the claim that the Patient-related questions contribute to the information status of the Patient comes from the use of pronouns as opposed to full noun phrases to refer to the Patient protagonist in the story. The short stories and the Patient-related questions each contribute to the overall discourse, with each of these contributions having their own topic. The first contribution in the discourse is the same as in Experiments 1 and 2: Information Status givenness introduces one protagonist before the other, driving the interpretation of the earlier one as the topic of the story.
Since the question is always Patient-related, the given protagonist of the task interaction is necessarily the Patient. The given protagonist of the first contribution can be the Agent or the Patient as established via the story. This gives two possible situations: in the first, the givenness of the Agent and Patient is somewhat balanced, with part of the discourse marking the Agent as given in the story and another part marking the Patient as given via the experimental task.
In the second instance, the Patient is consistently given in the discourse. This is in line with the finding that be-passives show no preference for one protagonist over the other and tend to report the whole event, including the Agentive by-phrase. This is also in line with the findings of Experiment 2, which indicated that get-passives occur more frequently when the Patient is marked as important.
It appears that get-passives still display an affinity for the Patient. When the Patient is both given and target of the question, it is unambiguously marked as the more important protagonist; when the Agent is given, yet the Patient is the target of the question, the attention, or importance is distributed between the two protagonists. In the former situation, with an important Patient, get-passives responses are more likely than when the Patient is not marked as important. When the Patient is framed as Patient-like i.