Is that word Feminine or Masculine?

Question

Write a program or function that takes in a single string containing only lowercase a-z, and prints or returns a truthy value if the word is the feminine version of the thing it represents and a falsy value if it is the masculine version. For example, hen is the feminine version for chicken and rooster is the masculine version, so hen might produce 1 and rooster might produce 0.

Doing this for all English words that reflect gender would of course be way too unwieldy. Your program/function only needs to support 20 masculine/feminine pairs. Below are five sets of 10 masculine/feminine pairs, categorized by topic. Choose any two of the sets; the 20 total pairs in these two sets are the 40 words your program/function must work for.

(format is masculine_version feminine_version)

1. General

   he she
   him her
   man woman
   boy girl
   male female
   masculine feminine
   guy gal
   lad lass
   mister miss
   sir madam
   

2. Familial

   father mother
   dad mom
   pa ma
   son daughter
   brother sister
   husband wife
   grandfather grandmother
   grandpa grandma
   uncle aunt
   nephew niece
   

3. Animal

   lion lioness
   rooster hen
   stallion mare
   bull cow
   drake duck
   boar sow
   buck doe
   ram ewe
   gander goose
   billy nanny
   

4. Royal

   king queen
   prince princess
   emperor empress
   duke duchess
   marquess marchioness
   earl countess
   baron baroness
   baronet baronetess
   lord lady
   knight dame
   

5. Fantastical

   wizard witch
   giant giantess
   incubus succubus
   nidorino nidorina
   nidoking nidoqueen
   ents entwives
   hanuvoite inimeite
   centaur centaurides
   merman mermaid
   khal khaleesi
   

So, for example, you might choose the General and Familial categories. Then any input from he to sir or father to nephew would produce a falsy value, and any input from she to madam or mother to niece would produce a truthy value.

The values don't all have to be the same truthy/falsy type, e.g. he might produce 0 but sir might produce false. You may assume only the 40 specific lowercase a-z words from your two selected categories are ever input.

The shortest answer in bytes wins. Tiebreaker is earlier post.

^{(This challenge is not meant to correlate with or make statements about any current gender-based social issues.)}

Answer 1

Retina, 26 bytes (sets 4, 5)

[^u]es|ee|m.i|y|^...c|d.*a

Retina is @MartinBüttner's regex language. I haven't used anything specific to .NET regexes, so you can test the regex at Regex101 here. Alternatively you can use Retina's grep G mode like so:

G`[^u]es|ee|m.i|y|^...c|d.*a

and pipe in a file with one word per line for batch testing.

Retina outputs the number of matches by default, giving us our truthy/falsy value. The rule "the values don't all have to be the same truthy/falsy type" is pretty important though since marchioness matches twice, giving an output of 2.

_{(Using the mod-chaining method from the previous male/female question seems to be shorter in CJam, but I'll let someone else do that)}

Answer 2

Retina, 39 32 bytes (sets 2, 4)

Accounting for marquess was annoying, since using ma was the best way to get some of the feminines.

[^u]es|ma$|mo|y|[mie]e|wi|ter|au

Try it here

Thanks to Sp3000 for his golf suggestion.

Answer 3

Retina, 28 bytes (sets 3,4)

w|[mhorse]e|[^u]es|duc|[dn]y

Retina uses .NET regex, but any flavor should work. Test it at Regex101.