How and when do these factors play a role in the process of spoken word recognition, according to the cohort model? Recognition is defined in the dictionary as “to know again from ones previous experience”. Spoken word recognition is an area of psychology that has had much research investigated into it, not only in recognising particular words printed and spoken but also how all the information that relates to a word becomes available. Many studies have been conducted and many theories have been produced to explain issues in spoken word recognition.
One Model that has been created is The Cohort Model that was proposed by Marslen-Wilson (1973,1975). This model is a localist model and was founded around the principle that as we hear words in speech, we set up a cohort of probable items the word could be. Then items are eradicated from this elected list and one is chosen to be the correct word. This model suggests that the cohort of probable items is triggered on the basis of approximately the first 100-150 ms of the speech input.
Through extensive research it has been revealed that numerous factors manipulate how easily a particular spoken word is recognised. One factor that is generally alleged to have a large impact on how easily a word is recognised is context effects. Context effects occupy a significant amount of information, for example it involves information ranging from previous sensory input to higher knowledge sources including lexical, syntactic, semantic and pragmatic information. For example it is believed that words, which are spoken in context, are more readily recognised than words that are not.
This is a wide area as it can include not only surrounding sounds and letters but also surrounding words and sentences that are also believed to have an impact on word recognition. Many priming studies have focused on the issue of how contextual factors affect word recognition. One experiment conducted which shows the importance of context effects on recognition was conducted by Grosjean in 1980′ which showed that participants needed a great deal longer on average to recognise a word which was out of context compared to a word which was within appropriate context.Context effects are also a large topic to discuss, as there are many sub-groups within it, which are discussed.
There is structural context which involves how words can be collected into higher level units and there is non-structural context, which refers to that which does not result in higher level representation and to how words can be combined into higher level units. In general according to the Cohort Model, context effects affect the pre-lexical stages of the recognition process as it affects the selection stage, which is before the access stage but after the integration stage.In more detail, structural and non-structural effects come into effect prior to recognition, which means that it happens before the ‘uniqueness point’ as this happens just before recognition. Fodor (1983) believed that this was the only type of context that could affect processes prior to recognition.
Structural effects are also sub grouped into different types. There is syntactic context that is the process by which a person almost predicts the subsequent word in a sentence by looking at the preceding words.This facilitates the recognition of words by looking at the context of the words surrounding the word needing to be recognised. There is also semantic context, which refers to word meanings and whether words, which are appropriate for the context, are recognised at a quicker rate.
The final one is interpretative context that involves pragmatic discourse and world knowledge. This is similar to syntactic context however instead at looking at the words preceding the word, which needs to be recognised, you use your previous knowledge of for example world events to facilitate recognition of the word.However the Cohort model was later changed and adapted. In this changed model, the amount of influence context effects have over recognition was reduced greatly and it was believed that instead of affecting selection and integration stages, it only affected the integration stage.
Irrelevant to the change in the model, one of the fundamental claims of this model is that when a word is heard in context in a sentence, sentential semantic constraints work together with the word recognition procedure to permit the suitable word to be selected faster than when the same word is heard in isolation.Another factor found to influence how a word is recognised is the meaning-level characteristics. It has been proven that the concreteness of a words meaning has been shown to influence lexical decisions, which means that concrete words have an advantage over less-concrete words. A study conducted by Balota, Ferraro and Connor in 1991 showed that words with concrete meanings are recognised more rapidly than those with more abstract meanings Another factor which is linked to the topic of meaning-level characteristics is the imageability of a word.
It is believed that words that are concrete have a larger and quicker recognition than words, which are abstract. This is believed to be because the person can “see” an image of the word and relate the word they are hearing to an image whereas this is not possible for abstract words. One of the Cohort Model’s principle claims is that there is early activation of multiple semantic representations. This means that very early on in the process of recognition of a word, does the meaning of the word have an effect.
Semantic priming is an additional influence over the recognition of spoken words, which is believed to have quite a large effect. This refers to the process by which identification of a word is made easier by having a word presented before it, which is analogous in connotation. Many studies have been investigated into this phenomenon and it is believed that the prime word facilitates the recognition of the second word however it has also been proven that the prime can delay the recognition of the target that is referred to as inhibition. Another type of priming is repetition priming.
This is the belief that once you have seen a word it is much easier to recognise it the next time. Although this is in the context of visual and not spoken word recognition it can be applied in the sense that once you hear a word once it is much easier to recognise it the next time you hear it. Frequency and related effects are a further topic discussed in this area of psychology. The idea behind this is simply that words that are more commonly used in speech are more readily recognised.
This topic was first discovered in 1951 by Howes and Solomon and has been carried on by many other psychologists.An important piece of research was done by Whaley in 1978′, which showed that frequency was the single most important determinant in the speed of recognition of words. Frequency effects include many other smaller issues that have been found to manipulate recognition of spoken words. For example age-of-acquisition, which is the age at which you learn at word has also been found to assist recognition and also the size of the word, as smaller length words are more common and therefore more likely to be recognised.
Another topic that ties in with frequency is familiarity; this is defined as a measure of personal frequency, which is something that is well known through long or close association. According to the Cohort Model frequency effects affect the activation level in the early stages of lexical access. It is suggested that there are relative frequency effects within the initial cohort. This insinuates that access in the cohort is not all-or-none but that it varies along a continuum.
Many experiments have been conducted into this area of research. One piece of research was carried out by Goldiamond and Hawkins 1958′.They found that frequency effects affect the later output processes and does not actually affect the recognition of the word, which means that frequency affects our response bias. According to the Cohort Model, frequency effects would come in effect in the post lexical stage after the integration stage as it is shown to affect the later response stages.
Word-non-word effects are another factor in the influence of recognition of spoken words. It was found in many studies that pseudwords are more likely to be remembered as they follow the formation of language codes and look similar to other words.However non-words such as “jehj” are much more likely to be rejected as they do not follow the formation of normal words. This idea is related more to visual word recognition, however its basic idea can be applied to spoken word recognition by saying that words that do not follow the formation of language codes when spoken are less easy to remember compared to words that do.
The Cohort Model accounts for this factor by suggesting that after the initial contact phase when words are being selected from the cohort of possibilities, the list of words are much shorter if a non-word has been heard compared to an actual word.This is because a non-word does not follow the rules of language and therefore factors such as imageability cannot help in identifying the correct word. One idea that has been discussed in relation to word recognition is the idea that there is an auditory input lexicon containing word recognition unit. This idea is based around the belief that every time a person hears a new word and gains knowledge of the associated meaning a new recognition unit is established in an auditory input lexicon.
After this, every time the word is heard, this unit becomes activated and the word becomes recognised.According to the Cohort Model this would happen before the integration stage, and would happen in the pre-lexical stages of recognition. This influences recognition because if the word is part of an already generated unit, then it is much easier to gain access to it, and therefore quicker to recognise. Word superiority effect is another influence over word recognition.
This is the belief that letters that are part of a familiar word are perceived faster than if for example the letter was on its own. Although this theory is referring mainly to written word recognition, it can be applied to spoken language byOverall there are many factors that are believed to have a large impact on the recognition of spoken words. There are also many other factors such as pronunciation of words and neighbour effects, which have however had less research investigated into them. The Cohort Model is a fascinating and clearly well defined theory on word recognition and is efficient in accounting for the significance of numerous topics ranging from word-initial information and also the rapid temporal nature of speech recognition.
However the Cohort Model fails to account for issues such as word-initial cohort errors whereas models such as the Trace Model takes account of this. The objective of language comprehension is not to comprehend individual independent words, but to understand the semantics of the words, as they are perceived in discourse. Much more research needs to be conducted into the area of word recognition in spoken language to fully understand how words are recognised and a new theory that includes issues collectively from many different theories might be necessary to understand word recognition completely.
