The difference between a phoneme and a phone is that phonemes are an abstract segment that are realised in actual speech as a variety of phones. This is where allophones come into it.
An allophone is a group of phones that belong to the same phoneme.
So, we have the phoneme - the smallest unit of sound that has meaning in a language. The phoneme is spoken, and this is the phone. However, many speakers produce phones that are phonetically different to the phoneme, but represent the same phoneme - these are allophones. For example:
The phoneme /p/ is in words like 'pin' and 'spin'. However, the phone in these words is different. /p/ in 'pin' is aspirated - [pʰ], /p/ in 'spin' is not aspirated - [p]. Therefore, the phoneme /p/ has (at least) two allophones - [p] and [pʰ].
I hope it is clear now. Tell me if it isn't.
A minimal pair/set is a pair/set of words that differ in only one phonetic element. So the words 'pin' and 'bin' differ only in the /p/ and /b/ phoneme at the beginning of the word. They are useful in linguistics because they show which sounds are phonemes and which aren't. For example, in English, the phone [p] and [pʰ] are allophones of /p/. This means that these phones do not create a new phoneme, thus they can be used interchangeably without a difference in meaning If I said [pɪn] or [pʰɪn], it wouldn't create a new word - the word would remain as 'pin'. However, /p/ and /b/ create a new word if they are changed. /p/ in 'pin' changed to /b/ creates a new word - bin.
This isn't the case in all languages, for example Thai. In Thai, /p/, /pʰ/, and /b/ are all phonemes. Thus changing one of these phones for the other creates a new word:
- ใบ /baɪ/ "sheet"
- ไป /paɪ/ "to go"
- ภัย /pʰaɪ/ "danger"
Onset, nucleus, and coda refer to the 'parts' of a syllable. The nucleus is needed in all languages, the onset and coda are optional - they may exist in a syllable, or they may not.
Think of the nucleus as the vowel. The onset is anything that comes before the vowel. The coda is anything that comes after the vowel. Vowels are represented as V, consonants as C.
So, the word 'a' in English has no onset and no coda. Shown as V.
The word 'be' has an onset and a nucleus, but no coda. Shown as CV.
The word 'at' has a nucleus and a coda, but no onset. Shown as VC.
The word 'cat' has all three - an onset, a nucleus, and a coda. Shown as CVC.
The onset and coda can consist of one consonant, or more - called a consonant cluster. In English, one syllable may consist of a maximum of 3 consonants in the onset, and 5 consonants in the coda. This is shown in the word 'strengths' (in some accents) - /ˈstrɛŋkθs/. This is shown as CCCVCCCCC. This branch of phonlogy that looks at how many phonemes can or cannot be used in one syllable of a language is called 'phonotactics'.
Let me know if you require any further guidance. I understand it's a bit much to take in.