Allophones of dental fricatives (/θ/, /ð/) in English

Question

I've noticed in my own speech (West Riding of Yorkshire, male, born in the '90s) two different ways I have of pronouncing phonemes /θ/ and /ð/:

1. The tip of my tongue sits in the gap between my top and bottom teeth. Air passes over it and is directed forwards or slightly upwards.
2. My tongue sits slightly behind my top front teeth. Air passes between the tip and my top gum and teeth, and is directed significantly downwards as it leaves my mouth.

I can tell the difference between these two methods of pronunciation by putting a finger on the outside of my top front teeth and seeing whether my tongue touches it. I think the rule is that I use method 1 word-initially and at the start of stressed syllables (e.g: “unthinkable”, “overthink”, “Athena”) and method 2 elsewhere.

My questions are:

1. Do these methods give rise to acoustically distinct phones? And are they thus allophones?
2. Is the distribution of allophones I described above common in English? Are there exceptions among English dialects?

Answer 1

It depends on what you mean by "allophone". Originally, "allophone" did not mean "the physical realization of a speech sound", it referred to a linguist's recording of speech as a set of phonetic symbols – a set of "phones". When the distribution of the symbols have been analyzed, you may find that certain set of phones have a predictable distribution, for example [t,tʰ,t̚] which gives rise to the concept phoneme, and the set of phones corresponding to a single phoneme are the allophones of that phoneme. Everything that you actually pronounce is the realization of some allophone of something.

The distinction between linguists' analytic construct and physical reality often gets blurred, so that now one might see actual physical realizations as being allophones, and again, everything is an allophone of something. The primary question, which we don't actually know the answer to, not about allophones, it is about whether the there is a perceptible difference. Not every difference in articulatory state creates a auditorily-perceptible difference.

Let's assume that we decide that you are producing [θ, ð] in some instances and [t̪, d̪] in other instances (also separately pronouncing [t, d] which are separate phonemes). Pending a final pinning-down of the context, you end up with classical complementary distribution between [t̪,θ] and [d̪, ð]. Whether we would say that [θ] is an allophone of /t̪/ or that [t̪] is an allophone of /θ/ depends on details that we don't have (given "ˈAthens" and "Aˈthenian", does the pronunciation of the consonant actually change?).