01-10-2025, 03:34 PM
(01-10-2025, 03:18 PM)Jorge_Stolfi wrote: You are not allowed to view links. Register or Login to view.(01-10-2025, 05:33 AM)Mark Knowles wrote: You are not allowed to view links. Register or Login to view.But what about when labels have the same word attached to distinct and quite different images
There aren't many such cases, are there?
People often use the same name for different things. Cancer in Latin could be a crab, a tumor, or a Zodiac constellation. The 28 Mansions (Lunar constellations) of the Chinese include 亢 kàng = "neck", 心 xīn = "heart", 胃 wèi = "stomach", and more...
The repeated labels could also be
- generic words like "tube", "reservoir", "dim", "bone", ...
- qualifiers like "big" and "little", "left" and "right", "upper" and "lower", ...
- misspellings of similar names, like "Hydrus"/"Hydra", "adductor"/"abductor"
- ordinal numbers, like "二" for 河鼓二 = "constellation River Drum star 2" = Altair
A problem also comes with the fact that many labels have spellings very similar to other distinct labels and can be found as repeated words in sentence text.
The idea that labels represent numerical references like "figure 45" sounds possible at first, but on inspection doesn't seem to fit.
It is on this basis that I concluded that some labels are null or meaningless. Given that these labels appear in the sentence text it follows that some words are null or meaningless. These words tend to be part of very repetitive spelling clusters. There are however words with quite distinctive spellings which I think are unlikely to be null or meaningless. So, I think words can be classified into two groups "filler words" and "real words". I am inclined to think that the majority of words are filler words.