For all practical purposes, for all the purposes which guide men's actions, public or private, the Russian and the Bulgarian, kinsmen so long parted, perhaps in very truth no natural kinsmen at all, are members of the same race, bound together by the common sentiment of race. They belong to the same race, exactly as an Englishman whose forefathers came into Britain fourteen hundred years back, and an Englishman whose forefathers came only one or two hundred years back, are like members of the same nation, bound together by a tie of common nationality. And now, having ruled that races and nations, though largely formed by the workings of an artificial law, are still real and living things, groups in which the idea of kindred is the idea around which everything has grown, how are we to define our races and our nations? How are we to mark them off one from the other? Bearing in mind the cautions and qualifications which have been already given, bearing in mind large classes of exceptions which will presently be spoken of, I say unhesitatingly that for practical purposes there is one test, and one only, and that that test is language. It is hardly needful to show that races and nations cannot be defined by the merely political arrangements which group men under various governments. For some purposes of ordinary language, for some purposes of ordinary politics, we are tempted, sometimes driven, to take this standard. And in some parts of the world, in our own western Europe for instance, nations and governments do, in a rough way, fairly answer to one another. And, in any case, political divisions are not without their influence on the formation of national divisions, while national divisions ought to have the greatest influence on political divisions. That is to say, _prima facie_ a nation and government should coincide. I say only _prima facie_; for this is assuredly no inflexible rule; there are often good reasons why it should be otherwise; only, whenever it is otherwise, there should be some good reason forthcoming. It might even be true that in no case did a government and a nation exactly coincide, and yet it would none the less be the rule that a government and a nation should coincide. That is to say, so far as a nation and a government coincide, we accept it as the natural state of things, and ask no question as to the cause. So far as they do not coincide, we mark the case as exceptional, by asking what is the cause. And by saying that a government and a nation should coincide we mean that, as far as possible, the boundaries of governments should be so laid out as to agree with the boundaries of nations. That is, we assume the nation as something already existing, something primary, to which the secondary arrangements of government should, as far as possible, conform. How then do we define the nation, which is, if there is no especial reason, to the contrary, to fix the limits of a government? Primarily, I say,