The document discusses different methods for measuring the productivity of morphological word formation rules, including type frequency, the ratio of actual to possible words, the number of newly coined words over time, and metrics involving hapax legomena or rare words containing the morphological pattern. Productivity is difficult to measure precisely due to the challenges in identifying all possible words and the occasional productivity of some patterns. The document evaluates several proposed measures of productivity and their appropriate contexts.
The document discusses different methods for measuring the productivity of morphological word formation rules, including type frequency, the ratio of actual to possible words, the number of newly coined words over time, and metrics involving hapax legomena or rare words containing the morphological pattern. Productivity is difficult to measure precisely due to the challenges in identifying all possible words and the occasional productivity of some patterns. The document evaluates several proposed measures of productivity and their appropriate contexts.
The document discusses different methods for measuring the productivity of morphological word formation rules, including type frequency, the ratio of actual to possible words, the number of newly coined words over time, and metrics involving hapax legomena or rare words containing the morphological pattern. Productivity is difficult to measure precisely due to the challenges in identifying all possible words and the occasional productivity of some patterns. The document evaluates several proposed measures of productivity and their appropriate contexts.
Productivity is a gradable property of morphological rules. Thus, for each rule we may want to ask how productive it is – i.e. we want to measure the degree of productivity of word- formation rules. Acceptability judgements Speakers tend to be more reluctant to accept new words than to accept new sentences, maybe because they do not encounter new words very often in ordinary life. Let’s take an example in the next slide. Consider the set of adjectives bearded (having a beard), winged (having wings), pimpled (having pimples), eyed (having eyes). The last word in this set, eyed, seems odd, and speakers may judge it unacceptable. But does that mean that it is truly ungrammatical – i.e. not allowed by the morphological system? People would judge ‘eyed’ to be odd because not all creatures have beards, wings and pimples, but virtually all have eyes, so one would rarely describe a person or an animal as eyed. Though, in different context, ‘eyed’ would be plausible, say in entomology. (I) The number of actual words (type frequency).
Productivity The concept of "type frequency" measures the
number of actual words formed following a specific measurments pattern in a language. It can be easily determined by examining a relible comprehensive dictionary. For example, the English suffix "-ment" has a high type frequency because there are many words like "investment" and "fulfilment" that follow this pattern, but it is not productive as it doesn't generate many new words. On the other hand, the suffix "- ese" may not have many existing words, but it is productive because it can be freely used to create new words that describe a specific language or jargon, like "journalese." (II) The number of possible words Measuring the number of possible words formed by a specific pattern is challenging, as it requires identifying all pattern restrictions accurately. Even then, the set of possible words doesn't necessarily reflect the likelihood of coining new words. There are unproductive rules without clear limitations, such as en-/em- prefixation in English, which should work with container-like nouns but it is not productive; because we have words like (entomb and embody), but we don’t have *embox or *encar. (III) The ratio of actual words to possible words (also called the degree of exhaustion). Is a challenging concept to measure. This is because counting all possible words is impractical, especially when it involves complex words formed productively. For example, in languages like English or German, N + N compounds can be created freely without restrictions, and the compound members themselves may also be compounds. This results in an almost infinite set of possible N + N compounds, making the degree of exhaustion for them low, even though many actual N + N compounds exist, and the pattern is highly productive. (IV) The number of neologisms (newly coined words) observed over a specific period, known as diachronic productivity, can be assessed with the help of a reliable historical dictionary like the OED. However, highly productive patterns may lead to new words being overlooked by lexicographers. An alternative approach involves analyzing large text corpora, such as newspapers from recent decades, to track the development and possibly decline in productivity of certain patterns, like the use of the suffix -gate in words like Watergate and Irangate. (V) The "category-conditioned degree of productivity" (P= V1,m/ Nm) This measure calculates the ratio of hapax legomena (words occurring only once in a corpus) with a specific pattern to the total token frequency of words with that pattern. This measure is based on the idea that productive morphological rules are likely to produce occasionalisms (rare words that occur infrequently), which is vital for assessing productivity. The formula for P is V1,m (number of hapax legomena with pattern m) divided by Nm (total token frequency of words with pattern m). It indicates the likelihood that a randomly chosen word with the relevant pattern will be a rare or occasional word. (VI) The " hapax-conditioned degree of productivity" (P*= V1,m/ V) This measure calculates the extent to which a specific morphological pattern contributes to the overall growth of vocabulary. It relies on the idea that productive rules tend to create rare, occasional words. In this measure, the number of hapax legomena (words occurring only once in a corpus) with a particular morphological pattern is divided by the total number of hapax legomena in the entire corpus, regardless of their patterns. This method allows the assessment of productivity at a specific moment in time without the need for a historical dictionary. These measures are most useful when comparing the relative productivity of patterns that serve similar functions, such as - ness vs. -ity in English. However, interpreting productivity measures can be challenging when considering different morphological patterns or when taken in isolation. Thanks for your attention For furher information and inquiries:
You can ask this: It’s NOT RECOMMENDED to ask this:
On the Evolution of Language
First Annual Report of the Bureau of Ethnology to the
Secretary of the Smithsonian Institution, 1879-80,
Government Printing Office, Washington, 1881, pages 1-16