Mini-grant project guide

LING  132: Language Processing . Summer 2025

Overview:

For the final project of this class, you have opportunity to propose your very own experiment, with the guidance of your instructor. Imagine that you are asking your department to give you $1,000 to fund a small study. How would you convince the funding committee that you have the necessary background knowledge and that the question and experiment is worth pursuing?  In an oral presentation you will describe the essentials ofthe proposed experiment, including:

• A clear statement of the problem with examples.

• A concrete hypothesis, along with viable alternatives.

• A prediction that supports the hypothesis.

• A method to test the prediction, and a justification for why this is an appropriate method.

Your presentation:

•  Should address parts 1 to 6 in the next page, and in the order displayed here. These parts must be fully developed for the presentation.

•  Should be 15 to 20 minutes long

•  I may ask you questions about the components of your presentation after you present.

•  You should present using slides, a handout, or other visual materials.

•  You should submit a copy of your presentation slides or handout on BruinLearn before you present.

Acceptable topics:

Unless you’ve discussed your topic with me, please concentrate on an area of language processing discussed in class, such as:

•  speech production and/or perception

•  lexical access

•  bilingual language processing

•  sentence processing

Unacceptable topics include: language acquisition, animal communication, computational approaches, or non-linguistic processing.  We haven’t explored these topics in detail in this course. Consult with me if you have any questions or concerns.

Oralpresentations: August6thto8th

**The short example proposal provided here is to serve as a guide.  The example shown here constitutes a short and incomplete mini-grant proposal. You should follow the instructions provided, and you should address all components of the instructions.

1 Introduction

A good introduction is important to orient your reader (me) to the goals of your project. Your introduction should include the following parts:

•  start with the main question.

•  highlight any crucial assumptions or previous findings that are directly relevant to the project. If previous findings contradict, explain how.

Note: previous findings can come directly from those discussed in lecture, if relevant. You do not need to look to outside sources in discussing previous findings.

•  state your hypothesis or hypotheses as concisely and concretely as possible.

•  state a prediction of your hypothesis, focusing on one in particular to test in your experiment.

•  summarize how your experiment tests the hypothesis and its predictions.

Short example

1. Introduction

Misperceptions in speech often occur when listeners encounter a sequence of sounds that is illegal in their native language. In a foundational study on misperceptions, Dupoux et al. (1999) find that Japanese listeners perceive an epenthetic and illusory vowel between the consonants of an illegal consonants cluster in their language even if that vowel is not present in the speech signal. For example, they found that Japanese listeners perceive non-words such as ebzo as ebuzo because bz is an illegal consonant sequence in Japanese. These effects have also been found for native English listeners (Berent et al. 2009; Davidson & Shaw 2012). In addition, other studies find that the lexicon also guides speech perception. In a finding known as the Ganong effect (Ganong, 1980), studies find that the perception of ambiguous sounds is warped towards the sound that form. a real word. For example, English listeners perceive an ambiguous sound between /b/ and /p/ as /b/ in the environment of -eef since beef is a real word of English but peef is not.

So, we know from previous findings that knowledge of phonotactics and the lexicon influence speech perception, but it remains an open question if phonological rules play a role. For example, given a language that repairs a certain illegal sequence in a very clear manner, do listeners misperceive the illegal sequence in a way that is consistent with their phonological rules? Or do they misperceive the sequence in some other way?

I hypothesize that phonological rules are accessed during speech perception. The central prediction of this hypothesis that is of interest in this paper is that listeners would be significantly more likely to perceptually repair an illegal sequence in a way that is consistent with their phonological rules. For example, in a phonetic categorization task where participants hear two nonwords and decide if the nonwords are the same word or different words, participants are expected to report “same” in cases where the repair item matches the native phonological rules of Korean for the given illegal cluster.

2 Design

You should carefully describe the proposed design with the following subsections. Each subsection should be about 1-3 paragraphs.

2.1 Participants. How many subjects would you run? How would they be recruited? What language they would speak natively? Highlight any other information about participants that matters.

2.2 Stimuli. Describe the kinds of stimuli your participants will be exposed to. Provide sample item(s) from your stimuli in a table or figure.

2.3 Method. Describe what the proposed task is: what participants need to do in the experiment, the information you will be collecting, and how. Explicitly state what the independent and dependent variables are, and what control variables you are including, if any. Inclue a table of figure that shows your variables with example experiment items.

2.4 Predictions. Elaborate on the possible range of results you might get from your study, and describe how each of these relates to your hypothesis. What kinds of results do you expect to see given what we learned in class, or given previous literature? What kinds of results support your hypothesis, and what kinds of results fail to support it?

Short example

2. Design

2.1 Participants. This study will test 30 native speakers of Korean. Participants will be recruited through UCLA’s SONA Subject Pool, therefore participants will be undergraduate students at UCLA. Participants will be able to take the experiment online.

2.2 Stimuli. Participants will be tested on Korean-like nonwords with a word-medial two-consonant cluster that is ilegal or illegal in Korean. The clusters we will test are the illegal stop-stop, stop-nasal, and fricative-stop clusters, and the legal /ns/ cluster. Korean shows alternations that repair the lattern three sequences at morpheme boundaries: the first stop of a stop-stop cluster alternates to agree in place of articulation with the second, the stop of the stop-nasal cluster alternates to a nasal consonant, and the fricative of fricative-stop clusters alternate to a stop. Some examples are shown below. The alternative repair assumed in this experiment is the vowel epenthesis repair. So Korean listeners will be tested on three kinds of items: items with the legal or illegal word-medial cluster, items with the phonological rule repair, and items with vowel epenthesis between the consonants of the cluster.

2.3 Method. Participants will be tested in an AX categorization design, frequently used in speech perception experiments. In a trial, participants listen to a sequence of two nonwords in a sequence. One of the nonwords contains the legal or illegal cluster, and the other shows either the phonological rule repair or the vowel epenthesis repair. Participants must decide if the two nonwords are the same word or different words. There will be a 500ms pause between the items in a trial, and the next trial begins once participants have entered their response. Participants can enter their responses by clicking “‘same” or “different” buttons on the screen. There will be a total of 100 trials in this experiment. The dependent variable, that which we measure, is the fequency at which participants respond “same” or “different”. The independent variables, those that we manipulate in the experiment, are the legality of the consonant cluster in Korean and the repairs (phonological rule or vowel epenthesis).

                                                      stop-stop        stop-nasal      fricative-stop           ns

phonological rule repair               pikka...pitka     temna...tepna     kakka...kaska             -

vowel epenthesis repair               pikita...pitka     tepina...tepna     kasika...kaska     tinisa...tinsa

2.4 Predictions. This study could have 4 different result patterns. The result pattern that would support the hypothesis is if participants report more “same” responses for illegal AX trials that show the phonological rule repair. This means that participants misperceive the item with the illegal cluster in a way that is consistent with their native phonological rules. The result that directly invalidates the hypothesis is if participants report more “same” responses for AX trials with the vowel epenthesis repair. Other possible results are if participants always report “same” or “different” regardless of the given repair. This would mean that participants are very accurately or inaccurately perceiving the items with no role from the phonological rules or phonetics.

3 General discussion

Give a brief summary of the central question and how the results from your experiment would bear on that question. You can discuss many things after this summary, but you must include a few of the following:

• Address any potential confounds in the experiment. How would you be able to tell that a particular variable was a confound?  How would you attempt to remove or control for such confound in a follow-up experiment?

•  Elaborate on other possible predictions your hypothesis makes.  Summarize an experiment you could run to put those predictions to the test.

•  Discuss other methodologies you could envision to test the same predictions of your hypothesis. How would an experiment under such methodology be conducted? What are the pros and cons of this new methodology compared to the one you proposed?

•  Speculate about what your hypothesis and predictions mean for our understanding of the language processing system as a whole. For example, does your hypothesis bear on any other areas of language processing beyond the one you’re testing for in your experiment?

•  Does your hypothesis and its predictions contradict those of previous literature? How are yours different, and what motivated you to argue the opposite?

For tips on finding relevant papers to you project, come see me! The optional readings for the course

are also a great opportunity to gather ideas, and a great resource for finding experiment designs.

If helpful, you can divide your discussion into multiple subsections.

4 Responsibility statement

If you are working on this project with others, state exactly what each of you contributed to the work. Group submissions without this statement will not be accepted. Final grades may be assigned individually  and not as a group. You do not need to discuss this part in your presentation, but please include this in  your slides.

5 References

Include any references you cited at the end of your presentation. I have no preference in format -just make sure your citation includes the full information. Here is a guide for citing inAPA: https://owl.english.purdue.edu/owl/resource/560 Below, you can find the citations for the papers I cited in the short examples above. This is the format generally used in linguistics.

6   Appendix of items

Finally, include a comprehensive list of example stimuli, organized by experiment design or conditions and dependent variables. Here, you can also include images that more concretely describe the experiment design and task (you can find these in the internet or in previous literature; make sure you include the source of the image). You do not need to discuss these in your presentation, but you should include them in your slides for future reference on my part.

**The short example proposal provided here is to serve as a guide.  The example shown here constitutes a short and incomplete mini-grant proposal. You should follow the instructions provided, and you should address all components of the instructions.