Background. The aim of the present study was to investigate whether word lexical and semantic properties may differently affect the timing and topographical distribution of ERP components. In particular, we focused on the neural processing of abstract and concrete words. Previous studies have provided controversial evidence about the way words having a different concreteness degree are represented in the brain. The reason of that may be traced, at least in part, in methodological heterogeneity and poor control of stimuli. Our efforts were directed to overcome such limitations. Methods. A group of 15 volunteers was engaged in a lexical decision task (word/non-word discrimination). 600 linguistic stimuli were created. They consisted of 300 words (150 abstract, 150 concrete) and 300 legal pseudo-words. All stimuli were balanced in terms of length. Abstract and concrete words were also balanced in terms of frequency of occurrence (Bertinetto, 2006) and familiarity, while they significantly differed in terms of concreteness and imageability (ratings were obtained from three independent groups of 30 subjects with 5-point scales). EEG was recorded from 128 scalp sites at a sampling rate of 512 Hz. ERPs were time-locked to stimulus onset. Low Resolution Electromagnetic Tomography (LORETA) was performed on ERP difference waves. Results. RTs to words were faster than RTs to pseudo-words (the so-called “word superiority effect”). Words were discriminated from pseudo-words since 300 ms post-stimulus with larger N2 responses to words than to pseudo-words over the left occipito-temporal areas, namely the BA37, possibly indexing the activity of the so-called VWFA. Similarly, at later processing stages (between 500-600 ms), corresponding to deeper linguistic processing and over the left temporal parietal electrode sites, ERP responses to words were larger than to pseudo-words. RTs to concrete words were slightly faster than to abstract words. The two categories were discriminated as early as 350 ms post-stimulus, with larger responses to concrete words than to abstract words over the medial occipital regions. Concreteness-related ERP differences were also observed in the amplitudes of the later anterior LP component (between 370-570 ms), with larger responses to abstract words than to concrete words. Conclusions. Overall, these data show how ERPs can dissociate between lexical and higher level processes. Our results indicate that semantic processing may take place near-simultaneously and in different brain regions with the processing of information about the form of a word and its lexical properties. The concreteness effect seems not to be strictly bound neither to some confounding linguistic properties of the stimuli, as the word frequency of occurrence or familiarity, nor to task demands. These data provide support for the hypothesis that concrete, imaginable concepts activate perceptually based representations not available to abstract concepts.