The implication of Whorf's position and similar views which deny a derivational relationship between zero-derived words and their bases are discussed by D. Kastovsky.
Found. Languagevol. 7 214
2015
It appears that zero-derived nouns account for a relatively small class of place nouns in present-day English.
B. Szymanek in P. O. Müller et al., Word-formation: Handbook Languages Europevol. II. vii. 1335
zero-derived typically occurs fewer than 0.01 times per million words in modern written English.
zero-derived is in frequency band 2, which contains words occurring between 0.001 and 0.01 times per million words in modern written English. More about OED's frequency bands
Frequency data is computed programmatically, and should be regarded as an estimate.
Frequency of zero-derived, adj., 1970–2010
* Occurrences per million words in written English
Historical frequency series are derived from Google Books Ngrams (version 2), a data set based on the Google Books corpus of several million books printed in English between 1500 and 2010.
The overall frequency for a given word is calculated by summing frequencies for the main form of the word, any plural or inflected forms, and any major spelling variations.
For sets of homographs (distinct entries that share the same word-form, e.g. mole, n.¹, mole, n.², mole, n.³, etc.), we have estimated the frequency of each homograph entry as a fraction of the total Ngrams frequency for the word-form. This may result in inaccuracies.
Smoothing has been applied to series for lower-frequency words, using a moving-average algorithm. This reduces short-term fluctuations, which may be produced by variability in the content of the Google Books corpus.
Decade
Frequency per million words
1970
0.0033
1980
0.0033
1990
0.0033
2000
0.0033
2010
0.0032
Originally published as part of the entry for zero, n. & adj.