(EDIT: question simplified)
This is a basic question, since I am new to data analysis and statistics. I am carrying out a 2X3 factorial survey where participants are presented with a series of sentences (varying with respect to which of 2 noun-types and 3 verb-types they contain) and are asked to make one of three choices for each sentence (T, F or C). There are 6 sentences of each of the 3 verb-types, with each sentence's being randomly assigned one of the 2 noun-types. So each participant receives 3 sentences in each condition. I am testing a number of hypotheses, including that: i) F is more likely than T for noun-type 1 with verb-type 1; ii) C is more likely for noun-type 1 with verb-type 1 than C is for noun-type 1 with verb-type 2; etc. The primary purpose is to test hypothesis i); the other hypotheses are intended to show contrasts and rule out some potential explanations if the data supports hypothesis i).
I am unsure how to analyse the resulting data. Since it is categorical, it seems I would I need to use non-parametric tests. But I'm not sure if it counts as independent samples (i.e. six different conditions) or repeated measures (because the same participants judged all the sentences). In the former case, I believe I need to use the Kruskal-Wallis test, and in the latter case, the Friedman test. On the other hand, perhaps I could take the proportions for each choice in each condition and use some parametric test; though I am not expecting a normal distribution. Any advice about which test to use would be greatly appreciated!