First comments to the first attempt

• One of the first things that is necessary to say is that the order between the external argument, the action and the internal argument mustn't be the one in there. If that were like that we will have to assume an exocentric system and a priori we don't want to assume that. We will think that the features are in a line from top to down, being in the top the features corresponding to the external argument, then the features corresponding to the action and then the features of the different arguments
• Other important thing is that even though in the paper the features are phrases, it's almost sure that the system will not have a specifier position. The phrases will just have a head of the feature
• It's important to stand out that movement is necessary and that movement will only by the phases and to an interphase position. The ending order of the different elements will be ruled by certain conditions
• Of course nothing is definitive yet and there are a lot of things that must be carefully studied. By now, the most important thing is to try to eliminate the miscellaneous phrases that the generative system has with categorial, functional, lexical... types. We just want a universal, ordered, homogeneus set of phrases which in this system will correspond to different semantic features. Any comment, critic, suggestion will be fundamental to develop this system