Yes, that is normal and expected. There are 3 classes of GO terms: cellular component, molecular function, biological process. Any particular gene can have any number of associated annotations in any of those. AT1G01200 (RABA3) is a nice example of that. Look it up in pubmed to get a better idea of why you're seeing this (this will also explain why this is normal).

Yes, it's normal. That's because GO is subdivided in 3 major components, each of them is subdivided and so on (i.e. http://respiratory-research.com/content/figures/1465-9921-12-63-4-l.jpg), therefore a gene can be classified under several terms.

Well adding to the notes of dpryan79 and JC: not just that the 3 major types of GO components- BUT also 'a single gene' can be involved in multiple functions, processes and locations ! For example- Cellulase- could be secreted to extracellular media, present in the periplasm, in the cytosol, in the vacuole or lysosome or in the chloroplast; function-wise it may be up-regulated as a response to pathogen infection, required for normal cellular metabolism, in xeno-bio-degradation of externally fed cellulose to a system, and could be involved in defense, homeostasis and thus spanning multiple pathways, networks, signalling events etc. ! That's why we have come to much more holistic approaches of GO analysis as GOSlim, GOFat terms etc. ! And yea, the GO annotations keeps expanding every year- depending on the robust annotation that is ever growing" as a result of both in silico (majority) and in vitro (mostly)/in vivo experimental results !

Yes it is normal, this is due to the fact that the structure of GO resemble a hierarchy where you can find parents and child nodes. In other words you can find general terms (higher levels of hierarchy) and more specific terms (lower levels of hierarchy) for a given gene or its product.

www.geneontology.org/GO.ontology.structure.shtml‎

If you want to visualise the percentage of shared gene between categories you can draw a heatmap like so