When John runs, he is, whatever else he is doing, transferring energy to the ground beneath him. But when we say ‘John runs’, it never seems to matter to us what is happening to the ground. We are just interested in John’s running. So ‘runs’ in ‘John runs’ is a one-place relation. John is the sole “central participant” in the situation comprising his running, at least when … (see below). RUNS(John).
When Joe eats, or Satish reads, sometimes all that matters to us is that Joe is eating, or that Satish is reading. In that case, the corresponding relations are the one-place relations given by EATS(Joe) and READS(Satish). But sometimes it does matter to us what Joe is eating, or what Satish is reading. In that case, different relations come into play, namely, the two-place relations given by EATS(Joe, salmon fillet with barcode 1123581321) (why I impishly specify the barcode may or may not become clear in later posts; it is something dba-related), READS(Satish, Die Phaenomenologie Des Geistes). What Joe is eating matters to us at the moment because he is discussing his plan to lose weight; what Satish is reading matters because we know that, given what he is reading, the wiring in his brain is in danger of becoming a tangled mess. Likewise, sometimes it matters to use where John is running to. John is running to a place where the tiger chasing him cannot easily turn him into a meal. RUNS(John, place where the tiger cannot reach him, identified by GPS coordinates 95°23’29″W, 29°48’27″N).
So EATS is not a single verb, because it sometimes names a one-place relation, and sometimes a two-place relations. Ditto READS and RUNS.
In Chapter 3, Some basic linguistic relations chapter, of their The syntax of natural language: An online introduction using the Trees program , Beatrice Santorini and Anthony Kroch distinguish between semantic arguments and syntactic arguments in an attempt to explain why these verbs sometimes take just one argument, sometimes two. Semantic arguments are the “central participants in a situation.” This is at least intuitively clear to me.
Syntactic arguments are…well, I am really not clear what a syntactic argument is supposed to be. “Syntactic arguments, on the other hand, are constituents that appear in particular syntactic positions (see Chapter 4 for further discussion)”, say Santorini and Kroch. Doubtlessly what a syntactic argument is as distinct from a semantic argument will be completely clear to me, in fact, as obvious as dust, when I have digested Chapter 4, and everything I am saying now will become clearly beside the point. But at the moment, I do not know what a syntactic argument is.
So for now, I will explain the ‘sometimes takes one argument, sometimes takes two arguments’ character of RUNS, EATS, and READS by claiming that EATS, for example, sometimes names a two-place relation and sometimes a one-place relation, depending upon the context. EATS, RUNS, READS are ambiguous.
I am throwing this explanation up into the air, with the intention of seeing how long or how short a time it takes for it to get shot down by someone with a better grasp of linguistics. Again, I am writing to learn.
Santorini and Kroch seem to think that, for example, EATS is a relation existing purely objectively, independently of contexts determined by what matters to us. If I try to specify as completely as I can what eating comprises, I will include the fact there is an eater as well as food that goes into the eater’s alimentary system. “eat denotes a relation between eaters and food.” On the other hand, my attempt at a complete specification of what running comprises would include the fact that there is a runner as well as energy that the runner is transmitting to the ground. Should I then say ‘runs denotes a relation between a runner and the ground’? No, because what counts as a ‘central participant in a situation’ depends upon what matters to us, not on an exhaustive description of the situation by some perspectiveless, omniscient being for whom all aspects are equally important. No mattering, no centrality. No mattering, no relations named by verbs. Different aspects of a situation that matter at a given time, different relations named by the same word, e.g., EATS, RUNS, READS.
That EATS, for example, may name different relations depending on the number of arguments it takes may get obscured a bit by the fact the flow of conversation can easily turn, at any moment, towards the topic of what Joe is eating. That something is getting eaten is always very close to the surface when we say “Joe eats”, especially given that the flow of conversation can at any moment very easily turn towards the topic of what Joe is eating. It can easily become something that matters to us, requiring a different relation. On the other hand, the flow of conversation never seems in danger of turning towards the energy that the ground is receiving when John runs. So we never seem to need RUNS to name, in addition to RUNS(actor) and RUNS(actor, goal), the additional relation RUNS(actor, object getting transformed by receiving energy, goal). Contrast this to THROWS, where what gets transformed by the reception of the throwing energy does usually matter to us. THROWS( Travis, ball with barcode 1235, Tinh). Travis throws the ball to Tinh.