Strange match span

[dishmint]

I don't understand why coolcar.cool was matched as an adjective.

In[107]:= 
LexicalCases["coolcar.cool car", 
  TypeToken["Adjective"] ~~ " car"]["Data"]

Out[107]= {<|"Match" -> "coolcar.cool car", "Position" -> {{1, 16}}|>}

[dishmint]

This means the model is not identifying canonical Adjectives, but inferring them.

[dishmint]

In[9]:= ExpandPattern["coolcar.cool car", TypeToken["Adjective"]] 

Out[9]= "coolcar.cool"

In[10]:= ExpandPattern["coolcar cool car", TypeToken["Adjective"]]

Out[10]= "cool"