Investigating the regularities underlying the linguistic use of metaphor is investigating the metaphor that allows speakers to use, for example, the sentence blew up with to mean Mary became enraged with me, not Mary's body did explode. task of describing specific knowledge about linguistic metaphors (LM) is motivated by (i) cognitive studies on metaphor, which analyze LMs as stretches of language as instantiations of conceptual metaphors (CM) (Lakoff & Johnson, 1980), i.e., particular activations of cross-domain mappings between the S(ource)-domain and the T(arget)-domain, and by (ii) natural language processing studies on metaphor representation, which assert that successful natural language understanding by computers depends strongly on explicit knowledge representation (Martin 1990). In short, if the LM Marry blew up with me instantiates the conceptual metaphor ANGER IS CONTAINER UNDER PRESSURE (Kvecses 2007), the issue to be computationally tackled is how to model this instantiation relation. As metaphoricity goes beyond linguistic dimension and pervades conceptual and perceptual dimensions, specifiable structures from these dimensions, such as conceptual domains (Lakoff & Johnson 1980) and image-schemas (Johnson 1987), respectively, are used as cognitive-linguistic constructs for describing this instantiation relation. These constructs are complemented by the analysis of the computational linguistic constructs semantic fields/frames (Nerlich & Clark 2000). This investigation includes the analysis of the vehicle blow up/explode in the domains of INTENSE EMOTION, COST, INTENSITY, and QUANTITY (e.g. My dad blew up when he saw the phone bill.; Gas prices blew up.; The birth rate exploded.; My head is exploding with fever.).We argue that the interpretation of the aforementioned LMs derives from two metaphorical inferences of the CM INTENSITY IS PRESSURE: (Inf.1) INTENSE X PRODUCES PRESSURE ON Y, and (Inf.2) WHEN Y BECOMES TOO INTENSE, Y EXPLODES. In turn, these inferences are instantiated by the following propositions:(P1) causes(Intense_X,Pressure_in_Y), and(P2) causes(Intense_Pressure_in_Y,Y to Explode), where the variables X and Y are placeholders for elements from the S-domain and the T-domain, respectively. In order to describe the interpretation of the exemplified conventionalized LMs in terms of the inferences Inf.1 and Inf.2, we discuss the following correlations: 1. S-domain and T-domain; 2. S-domain and image-schema; 3. Metaphor inference and argument structure; 4.T-domain, argument structure, and the metaphor interpretation; 5. Cluster of vehicles and argument structure.Two generalizations can be drawn: (i) the lexical dimension of conventionalized LMs can be described in terms of conventional image-schematic and conceptual structures, and (ii) this description is rule-governed.