Task $3$: Copyright

Copyright, and especially eligibility for copyright, poses a particular set of concerns to dictionary publishers.

In this task, you will explore some of the issues related to using dictionaries as sources of linguistic data.

Consider the following excerpt, taken from the Collins COBUILD New Student's Dictionary. Which

parts of the excerpt would you suspect to fall under copyright, and which ones would you assume are

not protected? Mark the parts $($e$.$g$.$ in two different colours$),$ and explain your reasoning.

flattered $/\text{'}$fl$$t$$d$/.$ ADJ AFTER LNNK$.$v If you are flat$-$

tered by something that has happened, you are

pleased about it because it makes you feel im$-$

portant. I am flattered that they should be so

supportive.

flattering $/\text{'}$fl$$t$$r$$y$/.1$ ADJ If something is flat$-$

tering, it makes you appear more attractive. Her

hair has been cut in a shorter, more flattering

style. $2$ ADJ If someone's remarks or behaviour

are flattering, they are pleasing because they

show that the person has a high opinion of you.

It is always flattering to be asked to judge.

flaunt $/$flo:nt$/$ flaunts, flaunting, flaunted.

VERB If you say that someone flaunts their pos$-$

sessions or qualities, you mean that they display

them in a very obvious way in$.$ They drove

around in Rolls$-$Royces, openly flaunting their

wealth.

If linguists want to use the contents of the dictionary as a linguistic resource, they must consider issues

related to licence agreements. One of the approaches that tackle this issue builds on the concept of

masking. Using masking, we can modify the text collection in a way that does not interfere with the

linguistic phenomena we are interested in$,$ while changing the original source to such an extent that

the original licensing restrictions do not apply any more to the modified data. One example would

be to map each word type to a randomly generated string such that the word length as well as the

distribution and positioning of vowels and consonants remain the same.

Taking the same excerpt as before, consider the following attempt at masking the original data: