I understand that Coepi is a defective verb with no present, imperfect or future form. It’s given in 501 Latin Verbs as: Coepio, Coepisse, Coeptum.
Reviewing my vocabulary from Orberg’s LLPSI Cap 32 I came across what must be an error on my part as I had written: coepio, coepere, coepi, coeptus (III) - which I can find in Whitakers Words. The full entry in Whitaker’s Words is below: note that he doesn’t even give Coepio, Coepisse, Coeptum which is odd as I believe that this is the form that’s used. Can anyone advise?
coep.i V 3 1 PRES PASSIVE INF 0 X
coep.i V 3 1 PERF ACTIVE IND 1 S
coepio, coepere, coepi, coeptus V [XXXAO]
begin, commence, initiate; set foot on; (usu. PERF PASS w/PASS INF; PRES early)
coep.i V 3 1 PRES PASSIVE INF 0 X
coepio, coepere, -, - V [AXXEO] Archaic uncommon
begin, commence, initiate; (rare early form, usu. shows only PERFDEF);
The Whitaker entry also says “(usu. PERF PASS w/PASS INF; PRES early)” which is quite important.
Whitaker can be a bit blunt/misleading tool for a learner who hasn’t mastered all the basics yet (it can contain automatically paradigms that are not present in the actual Latin, although here they at least mention that the present tense is early, as you can read)… it’s rather a very neat tool later on for reading to give you all the morphological options, if your interpretation of the sentence doesn’t make any sense and some option didn’t occur you (and in this sense it is much faster, more effective and easier to use than Perseus or the Wiktionary).
All you have to know is that a positive Whitaker lookup result doesn’t make the form automatically valid. (In 98% it does, but the 2% is important too). It makes it valid only “theoretically” (= if everything was perfectly regular and simple, which it is not), but can deviate from the actual use.
The Perseus word-lookup, on the other hand, does not show what is not present in the Latin corpus. However even some non-defective forms are sometimes by chance not present in the corpus and yet it doesn’t mean they didn’t exist. (= you can’t have all existing and valid forms present in the corpus unless the corpus is really really huge…)