One very important concept in economics is 'goods'. They're one of the most fundamental building blocks of the tangible side of economics. If economics is about resource allocation, 'goods' are what the resources are being allocated towards. What is a 'good'?
It's not quite so simple as just a thing, like a TV or a sandwich; 'goods' are anything - any thing - that one can allocate resources towards that contains value for that allocator (rather confusingly, 'services' are 'goods' too, by this definition; so is something like 'charitable contributions'). Yes, we're going to have to plunge into the pool of abstraction. First of all, we have to think about time and space: is a hot coffee in the morning in winter in New England the same as a hot coffee in the afternoon at a swim-up bar in the summer in the Caribbean? Doubtful.
We have two broad strategies, as economists, for dealing with that problem. One might be to say, well, people have different preferences at different places, at different times. That's difficult to work with because we have to try and hit a moving target, so to speak, and since we can't know preferences for sure anyway we sure can't know changes in them either. A second might be to say, those two coffees are different goods. That's also difficult to work with because then it's difficult to compare things at all.
In fact, in either case it will be very difficult to generalize. On the individual level, knowing how you allocated your resources in this situation at that time doesn't help me describe what you did or predict what you'll do in the future: if I see what you did at 11a.m. on Thursday when faced with the decision of coffee versus tea, how can that help me figure out what you might do at 10a.m. on Friday when faced with the same decision. These aren't necessarily the same 'goods' at both times, and in any case, the difference might not just be the time or something else that I can measure; it might be your mood or whether you're especially tired or just feel like a nice cup of tea for some mystical reason.
Seems trivial, but raise that to the level of the market for coffee, or the global coffee industry, or the impact of consumer decisions on the American economy, and the difficulty of defining a 'good' has snowballed into modeling chaos. For example, it's impossible to properly think about the current climate in the market for oil without thinking of the market for oil in the future. Of course, in the real world, the financial world, it's well understood that things vary in space and time: that's why we have things like futures markets, which let you buy 'thing X at time Y' (which is just a special case of 'good X now', really). These things are considered separate (though connected) markets, with separate prices.
And that's just how they were treated by economists as we developed microeconomics. The definition of a good was allowed to be very, very flexible and abstract, so that these 'things' don't just vary in physicality but in time, space, functionality, and so on and so forth.
It's not just time and space, though. An example: think of a college education. Is the good being sold by universities a 'college education'? Is it a 'degree from college X'? Is it a 'college education of quality Y from college X'? The signaling model by Michael Spence wondered (and I obviously paraphrase wildly) if people would still pay for a college education if it the education was intrinsically worthless but had the value of 'signaling' that you were willing to give up four years to prove that you were great.
Should it be surprising that the cost of a college education has been rising despite there being a bigger supply of colleges? Maybe not, if our definition of the 'good' includes 'quality', perceived or actual; it's easy to build a new college, but impossible to build a new college with a reputation to rival Oxbridge or the Ivy League. The supply of that good, whatever it is, is fairly well fixed. Econ 101 is obsessed with 'supply and demand' analysis; the fact is, supply and demand analysis takes you very, very far if you're prepared to speculate properly on what a 'good' is.
This is all quite similar to the corporate strategy mantra of identifying your 'core competencies' and defining the industry. For example: railroads and train companies aren't 'the railroad industry', but 'the transport industry', competing with airlines and buses, and maybe even 'the food industry' if they serve food, etc etc. That leads us dangerously close to the kind of 'provider of transport services' corporate-speak euphemisms that plague so many firms, but it's pretty much the same question as how to define a good in economics.