So, the language acquisition process that governs a child's learning is simply:
1 – lexicalised items (words with strong meaning)
“mama” “no” “doggy” “food” “eat”
2 – super-structure (arranging words according to the environmental norm, with English its Subject-Verb-Object)
“mama, doggy no eat food.”
3 – cohesive sentences (lots of language input leads to a mimicry of accurately constructed sentences)
“Ma, the doggy isn't eating his food.”
4 – cohesive devices joining super-items (cohesive sentences joined by linking devices)
“Ma, the doggy isn't eating his food but he's eating my food”
(at this point, more mimicry leads to more complex structures)
5 – continual, cohesive and coherent speaking turns/written texts (whether written or spoken, the learner can now engage for continuous periods of time linking sentence frames together with contexual, lexicalised items)
“Ma, the dog is not only eating my food, but he is also ignoring his food. I thought you said that you had him trained? Is this what we can expect from all the animals in this house?”
As you know, children can't speak a language initially and therefore this process is seemingly sequencial. However with adults this is not the case as they already have a mother tongue. Adults learning a second language will learn grammar before they have sufficient vocabulary (lexicalised items). They will also impose their mother tongue's super-structure on the language they are trying to learn. Many other small melds occur which blur this process. This is not an issue provided the learner can identify the language items they are learning into these segments and emphasis each one as appropriate.
My issue at the moment is that I have been selecting and memorising lexicalised items, but I have found that: a) I already know many words which I hadn't realised (listening to catchy songs in Spanish means I recognise a lot of language, rendering many to-be-memorised items obsolete) and b) I have been using the super-structure of subject-verb-object (with slight modifications to suit the target language) meaning that I am brokenly conversational (have been conversing with quite a few people including the laundry lady who spoke about her children for half an hour).
This initial 'problem' provides an opportunity to accelerate the process by doubling up the lexicalised items with accurate expressions encountered in every day life. This should add a significant boost to the 'broken conversation' skills and more contextualised items mean stronger memory as well as very commonly used structures providing the first steps to the cohesive sentences.