Examining Online Syntactic Processing of Spoken Complex Sentences in Chinese Using Dual-Modal Interference Tasks

The DMI experimental paradigm can be used to examine how the working memory is involved in the complex syntactic processing. And to further assess the processing assymetry of the Chinese writing courses. Combining DMI paradigm with the last co-decision task and on the Matco moving window technique is a novel approach for measuring the cross of a spoken sentence processing in real time.

Demonstrating the procedure with me will be Peihan Chen, a college student at NTIN. Before beginning the experiment, select a total of 48 bi-syllabic Chinese words, including 24 words and 24 non-words for the Lexical decision task. Select the words from the Sinica Corpus Technical Report, taking care that all of the target words are of a medium frequency.

Create the 24 non-words by using two monosyllabic characters that are individually meaningful but whose combination is semantically anomalus avoiding bisyllabic character words with identical radicals. Manually collocate the 24 non-words in filler sentences and the 24 words in target relative clause sentences. Compose the auditory stimuli into 72 sentences including 24 subject-gapped relative clause sentences, 24 object-gapped relative clause sentences, and 24 filler sentences.

Then divide the 48 relative clause sentences evenly into two groups to create an incomplete counterbalanced design to forming 48 trials. To set up the software for the experiment, open an appropriate behavioral research software program and use the software to randomize all of the stimuli. Then configure the software system to record the response time, the accuracy rate of the participant's responses in the lexical decision test, and the post sentence comprehension based on the participant's keyboard inputs.

Include feedback regarding the participant's incorrect lexical decision or lack of response to be displayed on the monitor screen immediately after the participant's incorrect or missing response. Before beginning the dual-modal intrasentential interference task experiment, seat the participant in front of a computer and equip them with headphones. Provide the participant with instructions both in written form on the computer screen and in verbal form by the experimenter.

Instruct the participant to listen to the sentences played through their headphones while simultaneously performing a lexical decision task. Instruct the participant to decide whether the interfering visual probe displayed on the screen is a word or non-word and to press yes for a word and no for a non-word as quickly and accurately as possible. Inform the participant that a comprehension question will follow immediately after the sentence and remind the participant to listen attentively to the auditory sentence while simultaneously performing the task.

After giving all of the instructions, have the participant complete a practice section involving trials with feedback. After the practice session, start the dual-modal intrasentential lexical decision test interference task allowing the participant to take a break between every 24 trials. For a dual-modal extrasentential interference task experiment, compose the auditory stimuli into 96 sentences including 24 subject-gapped relative clause sentences, 24 object-gapped relative clause sentences, and 48 filler sentences.

Divide the 48 RC sentences evenly into two groups to created an incomplete counterbalanced design with 96 total trials with zero, three, five-digit combinations with each zero, three, or five-digit load evenly assigned to all of the sentence trials. In the experimental software program, randomly assign the participant to one of the two stimuli sets representing the combinations of two within subject factors of sentence type and memory load. Instruct the participant that they will need to listen to the sentences segmented into words and played through the headphones and to pace themselves as quickly as possible by pressing the keyboard to initiate the playing of the subsequent segmented word.

Inform the participant that a question mark will appear on the screen that is followed by a yes no comprehension question related to the information they will have just heard in the preceding sentence and that they should answer the question. A short beep sound will be played when the participant presses the yes or no key to answer the comprehension question. Then, instruct the participant that after the beep the participant should follow the instructions that appear on the screen and should repeat the digit they observed prior to listening to the sentence.

To start the experiment, provide the participant with the 1500 millisecond visual representation of the digits. Then start the self-paced listening auditory moving window task and instruct the participant to keep the preceding visual presentation in memory. Because stage recall induces a greater memory load, instruct the participants to prioritize attempting to remember digits correctly while listening to the sentences as quickly as possible.

While the participant is completing the task, record the participant's digit recall responses on a score sheet. In this representative intrasentential experiment the ORC sentence type exhibited a processing advantage at the post matrix subject after the relative clause and at the end of the sentence. Whereas the SRC sentence had an advantage at the post matrix verb after the relative clause.

The simple main effect of sentence type was significant at the post matrix verb region and the post sentence region indicating that the SRC sentence had a lower processing load with the lexical decision interfering task at the matrix verbs. While the ORC sentences had a lower processing load at the ends of the sentences. Under a zero-digit load extrasentential interference, participants showed an advantage for ORC sentences at the initial processing region of the relative clause that demonstrated an SRC processing advantage at the relative clause marker DE and the matrix verb regions.

Under the three-digit load extrasentential interference, no subject or object-gapped relative clause sentence differences were observed around the pre DE regions. However, participants demonstrated an SRC sentence advantage at DE and an ORC sentence advantage at the matrix verb region. Under the five-digit load extrasentential interference, participants showed an overall ORC sentence advantage around the pre DE and post DE regions in the matrix subject region.

Our results show that the Chinese RC processing demonstrate a dynamic pattern that leads the conjecture that no intrinsic processing asymmetry exists in the processing of Chinese RCs. The current DMI technique can be extended to the self-paced reading paradigm to investigate how working memory is involved in the processing of reading in a natural reading context with interference.