My question is more about the pronunciation of words like arbitrary, oblivion and experience. The way theses words are divided into syllables make them harder for me to pronounce them due to the lack of certain consonant clusters in my native language. Take for example the word "arbitrary"; according to dictionaries, one should pronounce this word like /ˈär. bi. tre.rē/. Now it is much easier for me to say it like /ˈär. bit. rerē/. Or instead /əˈblivēən/ I would say /əb'livēən/. Likewise /ikˈspi(ə)rēəns/ becomes /iksˈpi(ə)rēəns/ I want to know if I say theses words the way I divided their sounds, will a native speaker pick up on that?
A syllable is typically divided thus:
There is an onset containing a consonant or consonant cluster, followed by a nucleus containing a vowel or diphthong, and then a coda containing a consonant or consonant cluster.
Every syllable needs a nucleus (e.g. the word "I"). The next most common is onset + nucleus ("buy"). Codas are less common. This generalization is true across all languages. Of course, there are exceptions where there's just a nucleus and coda ("at").
Now take a word like "obey". This is clearly two syllables. But we have to decide: does the /b/ belong in the onset of the second syllable, or the coda of the first?
Linguists have proposed that there is a maximal onset principle. This principle simply states that the onset must be as full as possible before we assign anything to the coda.
Hence, for "obey", the /b/ should first try to fit into the onset, which it can. Hence /oʊ.bej/
Now try a cluster, such as "extreme". /ɛkstrijm/ The /m/ must be in a coda since it's the last syllable. And the /r/ must be in the second syllable's onset.
What about the /t/? To answer this, we ask: is the cluster /tr/ possible? Yes, it is! We can easily prove this by finding a word like "tree" that begins with the cluster. So /tr/ is in the onset.
What about the /s/? Is the cluster /str/ possible? Yes, this is possible too. Consider "string".
Finally, the /k/. Is the cluster /kstr/ possible? No, it isn't. We can't find any words that begin that way. (It's even hypothesized to be impossible according to the theory of sonority.) Therefore, the onset is full, and the /k/ must go to the coda of the first syllable. Result: /ɛk.strijm/
(By the way, languages disagree about which clusters are possible. Spanish speakers typically cannot make /s/ + consonant clusters, resulting in names like "Esteban" instead of "Stephen". Meanwhile, French can begin syllables with /pt/, /pn/, /ps/, which English can't do. But in general, there's a lot of overlap between languages.)
So in short your syllable divisions would be considered wrong. It should be ar-bi-tra-ry, o-bli-vi-on, e[k-s]pe-ri-ence.
Does it matter? In some words, it's very subtle. For example, when trying out your version of "experience", I think I would find it very hard to pinpoint how long the pause should be between the /k/ and the /s/ and the /p/. In some words, however, it can have ripple effects on nearby vowels. I think one would need to hear a recording to tell you how noticeable it is.