when I pronounce the phrase "It was good" in a context like this one:
Person A: How was your day?
Person B: It was good.
I think that "was" is reduced to wəz (with a schwa sound). The only word that gets stressed is the adjective. Am I right?
[ ɪt wəz ˈgʊd ]
I also think in American English, the T in "it" is held when it occurs before another consonant. It's held in the throat. It's not a sharp sound with a puff of air. Am I right?
Yes, except "it" has some stress, I think: "2it was 1good." Yes, the t of "it" is held in the throat, in the sense that the glottis is closed during the t. A term for this is glottalization (though glottalization is also used for ejectives, which have an upward movement of the closed glottis).
p and k are also glottalized before another consonant (written p', t', k').
Sometimes glottalized p', t', k' are referred to as unreleased, meaning there is no audible puff of air at the end of the stop closure. The lack of a release is a consequence of the glottis being closed.
It's interesting to consider whether this glottalization affects the t part of the ch affricate (so-called). For me, it does, when the ch is at the end of a word.
For many people, the t' before a consonant can lose its oral closure, leaving behind only the glottal closure, which is a glottal stop. p' t' k' can also lose their oral closure and hence turn into a glottal stop before a stop consonant at the same place of articulation (not before s though, at least for me). In this case, the oral closure of the cluster is simply delayed.
ʼ (U+02BC MODIFIER LETTER APOSTROPHE) should not be used to refer to a consonant that is "merely" glottalized. It should only be used for stops involving full oral and glottal closure, larynx raising leading to increased intraoral pressure, and subsequent release of pressure with the release of the anterior (oral) closure.
̚ U+031A COMBINING LEFT ANGLE ABOVE.
sʼ, I forgot about that. But the voiced consonants in Yawelmani are described (by Newman (1937)) as glottalized, but not ejective. The point about unreleased seems to just be a misunderstanding.
I agree with @greg-lee that there is probably secondary stress on "it" in your example sentence. Regarding your second question about the t sound: you are correct that the /t/ in /ɪt wʌz gʊd/ is often realized "in the throat". The technical term for this in phonetics is glottalization, meaning that the glottis is constricted (partially or fully closed) instead of or in addition to the closure in the oral cavity. There are a variety of ways that linguists have represented this symbolically, for example:
"Glottalization" (or "glottalized consonants") is also a technical term in phonology, used to describe how sounds pattern together, so it is common to see these symbols used to mark phonological classes of sounds, in a way that does not necessarily reflect their phonetic realization in accordance with the definitions of the symbols in the International Phonetic Alphabet. For example, from a phonetician's perspective [tʼ] is technically incorrect in this case because it is reserved for marking ejective consonants1 which would not normally occur in that environment in English.2
Your point about the t sound in your example lacking a "puff of air" (called a release burst) is also correct. In this case, the most likely cause is that there was no oral closure to begin with (i.e., the glottal closure replaced the oral closure, rather than co-occurring with it). The lack of a release burst can also occur when full oral closure is made (in English this often happens with utterance-final stops, such as the p in "He wore a red cap"). In such cases, the buildup of intraoral pressure is usually released through the nose shortly after the end of the utterance, making subsequent release of the p closure silent.
1 Phonetically speaking, ejectives involve full closure of the glottis, complete or near-complete closure of the oral cavity (i.e., with the tongue or lips), a raising of the larynx to compress the air in the oral cavity, then a release of that pressure by release of the oral closure. By this definition, ejective sounds cannot be voiced (because they require complete glottal closure), and cannot be unreleased (because otherwise they are inaudible).
2 Similarly, the first symbol is phonetically inaccurate since the "combining tilde below" diacritic is defined as marking creaky voicing (hence it cannot be used with a voiceless sound like [t]).
