PRP4010 Language Production

PRP4010: Language Production: Learning, Meaning and Speech

Dissertation Assessment

Question: Dissertation Assessment

This assignment falls under “Dissertation Assessment” which was successfully solved by the assignment writing experts at Assignment Hippo Experts under assignment help service.

Dissertation Assessment

Case Scenario/ Task

  1. Timed picture naming norms provide a useful basis for future research and clinical stuff, and we don’t have them for UK English, so this project provides some.
  2. People have previously collected untimed written norms for these pictures in UK English.
  • How do our norms compare?
  • Do we have any evidence that timed norms are better than untimed written norms for predicting timed picture naming?

People have previously collected timed spoken norms for these pictures in US English.

  • How do our norms compare?
  • On the basis of these results, can US norms be useful in predicting names from UK participants? How would you use them most productively?

Previous norms have ignored semantic relationships between the pictures, and recent experiments suggest that we shouldn’t have to worry about them as long as we space related pictures at least 8 trials apart.

  • What (if any) evidence do we have that semantic relationships between the pictures can affect timed picture naming norms?
  • Do they affect naming latencies?
  • Do they affect name agreement, nonresponses, or which names people produce?

If we have evidence that long-lag semantic relationships affect picture-naming norms, then what does that tell us about the conditions that produce cumulative semantic interference (hint: consider Schnur’s 2014 interpretations)

Solution:

Abstract

This study presents timed picture norms for 525 pictures in British English. The International Picture Naming Project database already contains normed database for 500 black and white line drawings for more than seven languages, including American English and limited norms in British English. The aim of this study is to collect information in both dominant and non dominant names for the specific pictures used for various categories.  Most of the dominant names for British English will be consistent with those for American English, but some will differ and it will be useful to identify those differences and strength. In addition to obtaining the norms it tries to find a single strong factor responsible for Semantic interference that hinders the response time when the objects are from the same category even after many trials(Howard et al., 2006). This study will be of great use for the researchers in the field of language production, cognitive scientist, and patients with aphasia and to obtain norm in Welsh and other languages.  

Key words: Semantic interference, cognitive scientist, aphasia, hinders, dominant

Pictures are recognized easily and it involves the same cognitive functions compared to perceiving the same objects. The early stages of language production or development for children begin by visually seeing the objects or pictures. To represent an object there are two different domains one by pictures or real objects and the other by written words. Researchers have found a significant correlation in cognitive functions when presented reading a word and recognizing the picture of the object (Cattell, 1885). For an efficient language production one has to know to read the word, understand the meaning, reproduce orally as well as written form. It is same when a picture is produced, they have to see the picture, know its name, meaning, reproduce orally and written form. (chronometric analysis refer and write here).  A physical object has several names depending upon the nativity of the place, culture, usage etc. Cattell (1885) was the first person to identify that naming a picture took more time compared to reading a printed word. He further found that cognitive function involved in reading a word is automatic whereas it is voluntary in naming an object. Following Cattell’s findings, (Brown, 1995) did a research that if proper training given to name pictures will significantly reduce the time compared to reading the word. He failed in this study. (Fraisse, 1969) in his research found a unique difference between naming a picture and reading a word. There is only one option when we read a word, what is printed it has to be read; but to name a picture there are many options to express it; that is the reason for more time taken to respond a picture. He presented in his experiment that a symbol ‘O’ was reproduced as circle, zero and oh, each has different response time. These are the fundamental studies that lead to the important models for picture naming in various languages.

According to (Snodgrass & Vanderwart, 1980), pictures should be standardized in order to use it for the empirical studies. These studies are the stepping stone for our upcoming project “Learning, meaning and speech”. As mentioned in the earlier studies that there is a difference in processing speed between reading and word and naming a picture. They compared both processing speed in applying different activities and identified that it involved two types of memory that are semantic memory and episodic memory. Semantic memory is the general knowledge related to the use of language and it allows us to learn new concepts and ideas and equip with new information around us. Episodic memory, on the other hand, recollects and saves information on the specific situation or episode. To derive the difference between the pictures and words in accessing semantic memory the tasks like naming and categorization are used.  To register that information in episodic memory, the tasks involved are recognition and free recall are used. Several researchers (Cattell, 1886, Potter & Faulconer, 1975, Paivio’s 1978) conducted various research parameters to find the difference between the processing speed of reading a word and naming a picture. On picture naming task they also observed which characteristics affect picture naming, in most of the research work concluded that reaction time or the response produced after seeing the picture or word is inversely related to their frequencies. Words that are used more frequently have quick response time than the words that are used occasionally (Oldfield & Wingfield, 1965), the age of acquisition affect the frequency of naming latency (Morrison, Ellis, & Quinlan, 1992). Further, (Lachman, 1973) and his colleagues pointed that name agreement is another important features of reaction time in picture naming tasks. And the result of these studies pointed out that pictures have three important attributes that make it superior to words. And the three attributes are; pictures are dually coded, it has an image and each picture is unique on their own. Using this information their study focused on standardising 260 pictures. The pictures should be easily picturable, widely used norms and it should be represented at its basic level are the three important criteria used for the research work. Those 260 pictures are assessed on name agreement (the subjects have to identify each picture and writes its name), familiarity (rating scale -have to rate 1 as unfamiliar and 5 very familiar), visual complexity (rating scale – 1 as very simple and 5 as very complicated), image agreement (how pictures are represented  by their mental image- rating scale – 1 poor agreement and 5 high agreement). Thus, the standardized dominant names and alternative names of 260 pictures were obtained from this study (Snodgrass and Vanderwart, 1980). Naming latencies were found for these 260 standardized pictures and are used widely in picture naming studies (Snodgrass & Yuditsky, 1996). Thus, these norms were the fundamental base for the various psychological research studies, linguistic studies, norms in various languages, bilingual studies and latest fMRI studies on picture naming tasks.

The various research work has given an idea that the standardized normed pictures in English are limited and decided to include more items for pictures and used 520 pictures (Szekely, et al., 2003). Moreover, Snodgrass and his colleagues argued that, 260 pictures cannot be interpreted in one session by the participants and hence divided the list into two sets and conducted their research.  Szekely research disproved the concept and obtained fruitful norms for 520 pictures in a single attempt in almost 45 minutes from the participant. This study used the pictures from the previous studies and included further subsets of items and compared for validity and reliability from the prior studies. Validity is compared on (name agreement, alternate names, naming latency) and reliability of these pictures by presenting the pictures in different orders and proved it successful. The result of this study points out four output, First validating 520 pictures in a single session  against previous 260 pictures (Snodgrass). In older studies the response is written for the given picture whereas the response is elicited orally in the newer studies. The response time is shorter for frequently used words and longer for alternate names for the given list. The second point has they compared the results for old and new items, all the items were categorized into nine elements (people, animals, body parts, vehicles, foods, things to wear, small artifacts, large artifacts, nature objects). The study projected that the semantic categories and reaction times are different from old and new items. The third point focused on the effect of duration and the order of presentation, this was done by regression analysis. The results are there was a significant increase in naming latency and it does not affect the duration of the time period, but there was the change in the response time towards the end of the duration. The final point is to check on reliability, this was assessed on Cronbach’s alpha values and concluded that the reliability was high for this study. These picture norms are compiled in a database for researchers from various countries to develop the norms in their own cultural populations (norms in different languages), and it is used extensively in language studies, and for patients with aphasia.

Few drawbacks in these picture -naming tasks are; certain objects or items are not available after few years and the next generations are not aware of it. That is the familiarity of certain items is not same for years and years. For example, gramophone is not available in this generation and children are not aware of it. Thus, the variables keep on changing over a time period and new objects have to be introduced according to the present research studies. The variables have to be controlled by the researchers in order to carry out experiments successfully. The norms for the 520 pictures have collected for American English and not for British English (Bates et al., 2003; Szekely et al., 2004).  There are relatively few differences in American English and British English. To obtain British English norms, Johnson and his colleagues (Johnston, Kevin, Glyn W., & Chris, 2010)included 539 more pictures and decided to conduct research focusing on experience or frequency of the items and age of acquisition. The experiment had two parts and the participants were native English speakers from the UK. In the first part of the experiment, the clients recorded the rating for the variables (Age of acquisition, picture name agreement, familiarity and complexity) and are compiled into the database. Then in the second part of the experiment, the clients have to name the pictures in the given time. These results are compared with the previous pioneering studies of Snodgrass and Vanderwrt (1980), Morrison Chappell, and Ellis (1997) Barry et al. (1997), Szekely et al.(2003). The results are the age of acquisition and familiarity of the items are highly correlated when compared to the other studies. Picture name agreement correlates highly with Szekely et al.(2003) and less correlated with Snodgrass and Vanderwart (1980) norms.  And Complexity variable are highly correlated with Morrison et al., (1997) and less correlated on Szekely et al., (2003). The response time variable highly correlated with both Barry et al.(1997) and Szekely et al. (2003). The experiment proved that an individual can name more picture names in one slot without any discrepancy on the quality. Thus, these data are valid for further research (Johnson, 2010).

We learn words as soon we start talking and the vocabulary keeps on increasing throughout our life.  Remembering those words, comparing, analyzing, learning new words, reproducing spontaneously from memory are all the important tasks in language development. In the picture naming tasks, how the words are produced quickly and easily are quite challenging.  Each and every word we learn is linked to each other in specific categories and it either pushes or pulls other words and retrieve the correct match to the picture and reproduce it. As mentioned in earlier picture naming norms experiments, that words which are used frequently are retrieved easily from semantic memory. For example, in the animal category for ‘dog’, its relevant words like (pet, bone, bark, puppy, tail, bite) are stored. While retrieval process, how frequently those words are used in the past are associated with the response time.  When the words are frequently used it is retrieved quickly and for unfamiliar words the retrieval time is delayed (Vigliocco, Vinson, Damian, & Willem, 2002). If we show a picture of ‘cat’ it takes the time to retrieve, because it experiences the push and pull effect, because dog and cat are from the same semantic category. From the lexical experiments, it is proved that anything that blocks the retrieval of words is known as semantic interference. It is slower to retrieve names from the homogeneous condition (dog, cat, tiger, horse, camel) compared to mixed condition (paper, sofa, cup, star, truck) (Howard, Nickels, Colheart, & Cole-Virtue, 2006). He argued that shared activation, competitive selection and priming are the three process involved in lexical retrieval of cumulative semantic interference. To explain this, in order to reproduce word for the given picture, from memory its semantic related words also gets activated – (this is known as shared activation), then these related words are in the queue and the word which is relevant for the picture is retrieved –  (this is competitive selection). Once the relevant word is retrieved it behaves as a primming for future retrieval by making the word familiar. This is the fundamental concept in lexical retrieval in picture naming experiments. These are referred as the cumulative semantic interference that causes the response time slower in picture naming experiments (Howard et al., 2006). In 2010, Oppenheim (Oppenheim, Dell, & Schwartz, 2010) included the fourth process in lexical retrieval as activation-dependent selection time (more activated words should be responded quickly). These experiments gave us a clear idea that we have the potential to learn more new words and compete among themselves and reproduce according to the given stimulus. These literature studies reveal that cumulative semantic interference occurs due to the reproduction of target word from the semantic queue by using four lexical process shared activation, competitive selection, activation -dependent- selection time and priming in the post-semantic and pre-phonological stage in language production. Following this study (Schnur, 2014) had discussed the effect of an increase in response time and repeated trials when pictures are produced from the same semantic category. The results of Schnur (2014) are reproduced here “the cumulative semantic interference model must be modified by which learning vanishes due to experience but not time, to boost connection weight strength it remain stable for both old categories and new categories”.

Having these fundamental ideas from the pioneering experimenters, the present study focuses on obtaining simple picture naming norms in UK English and to obtain a single strong factor that hinders in semantic interference. It is clearly known from UCSD Centre for Research in Language has already provided norms for timed picture naming in seven different languages around the world. In the recent years, the International Picture Naming Project database has norms for dominant names for simple black and white line drawings for 520 common objects. It has norms for American English, but for British English, only limited norms are available. Keeping this as a base our project is to collect dominant names and alternate names for the picture in British English. Even though the participants are from the native British English speakers there will be an influence of American English for few pictures or items, this can occur due to the influence of media. That is most of the British people watch TV programs, movies that are telecasted from America. This can influence the usage of the words in Americanisms. Further, it will be observed that first, whether participants corrected to Americanisms immediately generalize that correction to naming new pictures.  Second, whether participants corrected to Americanisms store that correction to be recalled on a delayed test. Finally, to obtain whether that Americanized recall is item specific. The second part of the experiment is to find whether any single strong semantic interference is present for newly acquired words or it accumulates from the push and pull of the older words. We know many words and still learn new words, how these words are stored in the memory and when necessary how it recollects and reproduces is the fundamental concept in picture naming experiment. As mentioned from the previous studies words that are used more frequently are retrieved more quickly. If for example in picture naming task the homogeneous categories take response time higher compared to mixed categories of pictures. The response time is based on the lexical retrieval due to shared activation, competitive selection and priming of cumulative semantic interference. The study is to find a single strong parameter of cumulative semantic interference when a new set of words are presented. There are various methods to learn a new word. For example, in the category of animals, two new words are introduced ‘agouti’ and ‘capybara’. In the first method, the new words are introduced casually like cat, dog, horse as an animal category. In the second method, a special importance is given by focusing on the two words alone. The aim of this study is to find the difference in semantic interference in both types of learning and find the error rates. The next day the same picture naming task is given and ask the participant whether they remember the words as new one or same as old names from the category.

Hypotheses

  1. To obtain better norms for pictures in British English and to standardize in the database for future research.
  2. To obtain a single strong factor that hinders in picture naming task, that is a single factor responsible for cumulative semantic interference.

Experiment1

Subjects

100 participants will be recruited from the Bangor University School of Psychology Subject Pool for a 2-session experiment. Due to the nature of the experiment, we will specifically recruit natively monolingual British English speakers (no other languages before age 6), with normal or corrected-to-normal vision and hearing, and no known language disorders. The purpose will be to simply collect name agreement for the set of commonly used pictures.

Procedure

The English-language experiment will include two sessions of approximately one hour each. In the first session, each participant will complete a computerized naming task, with feedback. Approximately 526 images from the IPNP set will be presented, one at a time, in the middle of a screen. Participants will be instructed to name each picture as it appears, with their responses captured by a head-mounted microphone and digital audio recorder. Following each response, the dominant British English name for the image, as identified by a previous study will briefly appear below the image as feedback (e.g. kettle). For a small number of images in the final third of the list, the dominant American English name will be presented instead (e.g. shrimp instead of prawn); the proportion of British to American names will remain constant across all lists, but the assignment will be counterbalanced. Participants will be instructed that these are the dominant names that they should remember to use throughout the experiment. The second session will proceed much as the first, but without corrective feedback. As a baseline, some participants will instead complete a no-feedback single-session version of the experiment. The idea is to gather some basic norming data that could support the suitability of the norms from the main experiment.  

Experiment 2

Subjects

50 Bangor University students, recruited via SONA for course credit or printer credit, and compensated according to standard rates.  This is an English-language study so will recruit native monolingual English speakers.

Procedure

The procedure is to briefly explain to the participants on some novel (real) words, and train them on these and some old words from the same semantic categories. By the end of training they should be able to consistently name all of the pictures correctly. Then we will test them on a standard picture naming task, recording their speed (and accuracy) when naming all of the pictures several times. The next day we will bring them back and run the same naming test again with the same items, and finally ask them about each of the novel words to confirm that they really were novel for them. For each categories half will be words that participants should already know (e.g. CAT), and half will be words that participants should not yet know (e.g. AGOUTI). Since the main task is picture naming, we will have pictures (black and white line drawings) for each of these words. These will be imageable count nouns, like DOG and CAPYBARA and FORK and TAGINE. Day 1: This session will last about an hour. We will start with a series of card-sorting tasks (~40 minutes), where participants will be sorting flashcards, each with a picture on one side and it’s name on the other. These will include several well-known words and several novel words from each of several categories. That is how we will try to teach people to name the pictures by going from seeing an image to activating a meaning and then retrieving a word. Then we will test them on a picture-naming task (~20 minutes), naming each picture, one at a time, as it appears on a computer screen. The naming task and semantic interference study will be based on experiments of Johnston et al (2010) , Howard et al.(2006) and Schnur (2014)

Day 2: This session should last about half an hour. Participants will return to repeat the same picture-naming task. Then we will present each of the images and names, one at a time, to ask whether they already knew the word (that way we can make sure that the novel items really were novel). Foreseeable variation: The second day may not be necessary. If, after collecting an initial batch of data, we see sufficient evidence of cumulative semantic interference among novel items on Day 1, we may omit Day 2, moving the “Did you already know this word?” task to the end of Day 1.

Equipment

Experiments will be run in the Bangor Language Production Lab (019 Lloyd Building, still under construction). Naming tasks will use computerized experiment-running software. Naming latencies will be assessed with a custom-built voice key (still under construction). Vocal responses will be audio recorded to verify response- and voice key-accuracy offline.

Study Design

Trial lists will be created such that each participant will name each of the pictures once in each of several large multi-category cycles. Trial orders within the cycles will be planned according to a sequentially counterbalanced nested Latin square design, so that subsequent analyses can recover linearly independent effect estimates for 1. semantic interference from other same-category items from the same subset (e.g. novel mammals); and 2. semantic interference from other same-category items from the same subset (e.g. familiar mammals). Analyses: Single-trial-level data will be analyzed via linear mixed effects regression, to separately estimate 1.) semantic interference from other same-category items from the same subset (e.g. novel mammals); and 2.) semantic interference from other same-category items from the same subset (e.g. familiar mammals); for the familiar and novel items. The method has some advantages over traditional ANOVA, but is largely analogous. The actual statistical model will include a binomial main effect of novelty (novel vs familiar) and its interaction with each of the two linear effects just described.

Conclusion

In the present study we will obtain standardized timed norms for a large set of pictures in British English and a single strong factor that is responsible for cumulative semantic interference. A strong reason for cumulative semantic interference effect in picture naming task will be found out. Future implications from this study will be to carry out research work on finding norms for pictures in Welsh. These materials will be of great use for future researchers in the field of language production, cognitive science and aphasic patients.

This Dissertation Assessment sample was powered by the assignment writing experts of Assignment Hippo Experts. You can free download this Dissertation Assessment answer for reference. This solved Dissertation Assessment sample is only for reference purpose and not to be submitted to your university. For a fresh solution to this question, fill the form here and get our professional assignment help.

Want latest solution of this assignment