Factor three relates to ‘Diversity of users and situations of use’. A successful new word will be used by a range of people in a variety of situations. At level 0 a word will be extremely specific to a certain group of people. This could come in the form of a technical term or even a slang term from a close-knit group of individuals somewhere in the world. However, moving up to level 1, the word will have expanded its audience and be in use by outsiders even though its meaning will still be specific. Level 2 sees a word becoming used by an array of people throughout societies. Moreover, the specificity of the word has become less restricted and needs no explanation in everyday publications.
The fourth factor relates to ‘Generating new forms’ and whether a new word has the ability and practices generating other forms and meanings; the likelihood that it does increases its chance of success. As Metcalf says on a words ability to branch out: ‘Like a plant that is watered and fertilized and gets plenty of sunshine, a successful new word grows’ (Metcalf 2004: 159). If a word has a Generation level of 2 it simply means it has a variety of meanings. However Generation level 0 is related to Diversity level 0; both relate to a technical word that is used only within a single context and is therefore because of its restricted usage, not likely to generate other forms.
Finally, factor five relates to ‘Endurance of the concept’ and what the word stands for. If a word is introduced for a concept current and in the specific moment, which is highly likely to pass in a short space of time, the word will receive Endurance level of 0. The reason for this is because the word’s existence will be too brief to leave any lasting legacy, as it will have no use once the concept it was connected to has disappeared.
2.1 Identified Problems with the FUDGE Model
After analysing the FUDGE model as a method of predicting a new word’s success, some fundamental issues become noticeable. Metcalf appears to be extremely general in his methods of analysing new words. Although his five factors seem to be relevant, it is not believable that they are the only jointly sufficient features that determine a word’s success. Furthermore the generality of some of the levels within the factors hold problems for certain types of words. This refers to the fact that Metcalf states within the ‘Diversity of Situations’ category a word will remain at level 0 if it is only used within a very specific group of people. However, this implies that all words which appear only within a specific use of situations or context will stand a lesser chance of success. I disagree with this strongly as it suggests that technical jargon specific to a certain field for example, if following Metcalf’s analysis, would have to be placed at level 0 for this factor. This implies that the word has a reduced chance of success; however this is clearly not the case in all circumstances which will later be demonstrated by my own results. This appears to be a drawback of Metcalf’s model as it would, in these circumstances, give us a less accurate score and thus affect our overall prediction for a word. Although the FUDGE model gives us a basis for predicting the success of new words, we need to recognise that Metcalf over-generalises not only the factors themselves, but also the levels within the five factors. Therefore, this could be seen to affect the overall legibility of our output prediction of success for a word. However, if awareness is shown to these problems and taken into account when interpreting results, the FUDGE model still provides a good basis for further predictions.
I suggest that Metcalf could begin to rectify these issues by expanding on some of his ideas. The rationale behind his model is that it proves to be a simplistic way of predicting a new word’s success. However, if, as in the example above, the problems can be seen to affect the viability of the prediction, it is pointless. For that reason, if he expanded some of the ideas behind his factors and accounted for some of the problems, the method would possibly prove more effective. Subsequently, if Metcalf were to show awareness of some issues within the model, such as having to award technical jargon a Diversity level of 0, amendments could be made which could result in more accurate predictions.
3. Methodology
The idea behind the FUDGE model is that it gives us a way of measuring a new word and thus a way of predicting its success. I will respond to the model by using its factors to measure my own data set from the year 1910 against, and consequently look at how effective it claims to be. I will do this by collecting a data set, measuring it against each factor by giving the word a score of 0, 1 or 2 and then look further at the words which make up the data set to see if they are still being used in present day English. The results of whether the words are still in use today will then be correlated to the FUDGE scores which my words were awarded. This will then allow me to show whether the FUDGE model does in fact provide an effective way of predicting the success of a new word. I will also draw on some of the problems highlighted above in relation to my data set, and suggest ways in which these may have affected my results.
4. Data Collection
My initial data set consisted of one hundred words which were randomly selected from the Oxford English Dictionary online (OED) from the year 1910. To achieve a random sample, every third word was chosen from the list of entries. To attain a more concise data set that I could analyse in more detail, I then took a further random sample of forty out of my original one hundred words which I then measured against Metcalf’s FUDGE Model. To give them the most accurate score I could and hence improve the validity of my results; I also used the Time Magazine Corpus in relation with the OED online to try and gain a firmer understanding of frequency of use and the other factors. Deciding upon which corpus to use initially proved problematic as I needed the most representative to ensure I gained accurate results. This proved challenging because the largest available corpus; The Corpus of Contemporary American English (COCA), which has a database of over 450 million words, only holds work from 1990 onwards. Therefore this would have been of no help in finding the frequency of use of my words from 1910. As a result, I went on to look at The Times Magazine Corpus instead. Although this corpus did not begin until 1923 and does not hold as much data as COCA, it was the closest most representative match I had. While this meant the data I was searching was still that from thirteen years after the year of my data set, it still aided me, along with the OED online, in awarding the most accurate score I could for the factor of Frequency of Use.
I faced problems like the ones mentioned earlier when analysing my data set and giving scores for each factor. For example some words I found were specific to a certain field such as ‘manganaxinite’ which is a specific chemical formula. I found scoring words such as these problematic because I had to score them low in regards to the factors of Frequency of Use and Diversity of situations especially, because of the specific contexts they were only found within. As a result of this, my results and scores for particular words were therefore implying that I felt the word had a low chance of success. However, if I was to adhere to the principles and scoring system of Metcalf’s model, I had no choice but to award words such as ‘manganaxinite’ low scores for factors such as Frequency and Diversity because of its nature. This is an example in practice of the problem I mentioned above with Metcalf’s model. The system of scoring which he has enforced is problematic in some senses which again as mentioned above can affect the results provided and thus the overall success prediction.
4.1 Graph to show the amount of words at each overall FUDGE Score from 1-10
FUDGE Score
Total Number of Words
This graph shows the relationship between the total FUDGE score (X axis) represented by the red column and the total amount of words at that score (Y axis) represented by the blue column. It shows that from my data set, no words were awarded the top FUDGE scores of nine and ten. The graph shows that my data set failed to yield high FUDGE results which suggest many of the words from my data will not still be in use in present day English.
5. Discussion of Results
To gain a firmer understanding of whether my results were reliable, I searched for the words which formulated my data set in a corpus and a present day English dictionary, to gain an idea of whether they are still in use today and therefore, whether my scores were accurate.
To do this I used The Corpus of Contemporary American English (COCA) because, as mentioned above, it holds a large representative selection of data from 1990-2012. Therefore, it would give me the most accurate results in relation to whether the words which formed my data set were still in use. From this information I will then be able to see if my original FUDGE scores which I appointed in regards to the model outline by Metcalf, were in fact effective in predicting the future success of a word.
Out of the forty words which I assigned FUDGE scores to, thirteen were not accurate. Some had very low scores which would imply they were going to drop out of use sometime after 1910; however were still found in recent data according to the COCA corpus. For example ‘bildungsroman’ which was assigned a score of only 1 due to its restricted meaning and unobtrusiveness was found in fifty six entries from 1990-2010. On the other hand, words which were assigned a higher FUDGE score which suggests they would become a success, failed to yield results from the corpus. For example, ‘newspaperland’ which was assigned a score of seven failed to yield a single result from the corpus, suggesting it is no longer in use in present day English. Errors such as these imply the model may not be as reliable as initially thought. Moreover, although COCA does contain a huge amount of data and is the best representative mechanism for finding words in use, we must remember that simply because it is not found in the corpus does not necessarily mean the word has fallen from use completely; it may still be used in some circumstances.
On the other hand, twenty seven scores did correlate to the success predictions. For example, ‘pong’ was given a score of eight and was found in three hundred and eighteen entries in the corpus dating from 1990-2012. ‘Curiara’ was given a score of just one because of its nature, and was not found in any data from the COCA corpus. These are examples of the FUDGE model and its factors being used as an effective way of predicting the success of new words. It shows words which received a high score and hence predicted a good chance of success, such as ‘pong’ were still found to be in use today. Whereas on the other hand, those that were awarded low scores and therefore predicted to fall out of use soon aft, such as ‘curiara’ were not found in any recent data.
6. Conclusion
To conclude, Metcalf’s FUDGE Model (2004) offers a way of predicting the success of new words which are ever infiltrating our language. To test the effectiveness and reliability of the model, an original data set of one hundred words from the year 1910, later compressed to a more concise set of forty was chosen. The words which formed the data set were then analysed in relation to the model, and assigned an overall FUDGE score, with scores ranging from one to eight. To then test the reliability of the scores, the words were searched in the COCA corpus. It was found that out of forty, thirteen FUDGE scores were not accurate in that they did not correlate with the supposable success prediction from the model which the scores implied. Therefore the results indicate that, although the model for the most part works as a successful, effective way of predicting a new words success, it cannot always be deemed reliable. This may be due to the drawbacks it possesses in that some of its factors remain too general and cannot account for all word types.
To a certain extent, Metcalf’s FUDGE Model (2004) works as an effective way of predicting the success of new words. However its problems need to be taken into account when interpreting results which the model yields. It needs to be remembered the results may not be one hundred percent accurate, however can provide us with a firm idea, in most cases, of a success prediction. To further test the model’s effectiveness and gain a better understanding of whether the problems mentioned in section 2.1 pose a fundamental threat, data samples from a wider range of time periods could be analysed in relation to the model. This would hopefully allow us to gain a deeper understanding of its value of the FUDGE model in this field of new words in language.