 Hello again, I welcome you to another e-lecture of the series about constituent analysis. In the lecture, constituent tests, we discuss the tests and arguments that we can apply in order to find out how sentences can be structured hierarchically. The focus of this particular e-lecture will be on more general issues, that is how do we label the results of such a constituent analysis. But let's repeat first. In the e-lecture, constituent tests, we looked at the sentence, this boy will speak very slowly to that girl. A sentence which was taken from Andrew Radford's 1977 book, Transformational Syntax. And we decided that, for example, very slowly is a constituent. Let's quickly outline the reasons for our decision. We said that we can pre-pose it very slowly. This boy will speak to that girl, or that we can post-pose it. This boy will speak to that girl very slowly. We can use very slowly as a sentence fragment. How will he speak? Well, very slowly. And last but not least, we can use a pro-form, namely so, to replace very slowly. This boy will speak so, to that girl. Or take to that girl another constituent. Well, even though fewer tests apply here, we have reasons to believe that to that girl is a constituent. We can pre-pose it to that girl. This boy will speak very slowly. We can use it as a sentence fragment. To who will he speak? Well, to that girl. And then we have other tests, semantic tests, where two is directional and two, in many languages, assigns a particular case to other constituents. So with constituent tests of this kind, we can justify the hierarchical organization of sentences. And the method to work out such hierarchies is referred to as constituent analysis. And this is how it works. So let's apply the principles of constituent analysis to our sentence. Before we start, however, let me give you some advice. Since we are going to develop a sort of growing hierarchy, it is advisable to write down the sentence at the bottom of an empty sheet so that you have sufficient space above it. So let's move our sentence downwards. Now here we have all the space in order to develop our constituent hierarchy. In a first step, we associate each word with its own branch and a label on top. The labels are referred to as nodes. Since these nodes, which are represented as little deltas here, are lowest in the hierarchy, they are referred to as terminal nodes. Let us now replace the deltas by respective syntactic categories. This is a determiner, boy and noun. Will is an auxiliary verb, speak is a verb. Very is a degree adverb, slowly an adverb to a preposition that another determiner and girl is another noun. So this categorical or word class or part of speech analysis is always the first step. Now we are ready to build constituents. To do this, we have several theoretical possibilities. On the one hand, we could opt for a really flat structure. A structure which looks like this, where we have one constituent that is associated with all the other nodes, all the other elements that we have identified. So here, the branching node on top dominates all other terminal nodes. Since a flat structure misses many generalizations and does not show real hierarchies, modern syntactic theories all use a hierarchical structure with several intermediate constituents. Like this. So let's confine ourselves to that option too. And let's return to step one and start all over again. Now here is our categorical analysis with the terminal nodes. Let's now add the first set of constituents where, for example, our determiners build constituents with their nouns. And where the degree adverb, very, builds a constituent with slowly. So here we have the first three constituents. And at this point, we can already introduce some terminology. The result of constituent analysis is a tree diagram. Interestingly, in syntax, it is an upside down tree. Now in real life, trees grow from bottom to top. But in syntax, they do it the other way around. So we are building a tree which eventually will end up with its root. The lines in this tree diagram are referred to as branches, just like in a real tree. And the categories within a tree, they are referred to as the nodes. So we have the terminal nodes here. And then we have the new nodes which label our constituents. Depending on a position within a tree. And here we have just a shorthand version of a tree, a so-called local tree. Depending on their position in a tree, nodes can be classified using family relations. In syntax, however, they're exclusively female. And this is how it works. Branching nodes, that is, nodes that have at least two branches, are referred to as mother node. The nodes below that mother node are referred to as their daughter nodes. So here we have two daughters, the determiner and the noun. And the relationship between the daughters on the same level, that is, the daughters of the same mother, is just like in ordinary family relationships that of sisters. So determiner and noun are sisters. They are daughters of the same mother. So in syntax, we have mothers, daughters, and sisters. We do not have father, sons, and brothers. OK, but that's life. With these family labels, we can now easily address the relationships between the nodes in a tree. Returning to our sentence, we now need labels for our constituents. And the convention for these labels is simple. Take the head of a constituent, that is, the element, that is, for whatever reasons, more important than all the other elements, and make it the name patron of its mother node. So let's do that. Nouns, for example, like boy and girl, are heads. They determine the determiner in such a way that the determiner, for example, receives the same number. This boy or these boys, that girl or those girls. So their mother node is referred to as a noun phrase. The head is a noun, and so we get a noun phrase. Now take very slowly. The head is clearly the adverb, and not the degree adverb. The degree adverb just modifies the head. So we have something like very slowly, somewhat slowly, extremely slowly, and so on and so forth. As a result, we get an adverb phrase, where the adverb is now the name of that phrase, adverb phrase. Let us find more heads. Another head is two in to that girl. Now here the preposition is the head of the entire noun phrase, and consequently, we get a prepositional phrase as the common mother node. But why can such a small element be the head of a noun phrase? What does it do to the head? Well, not very much in English. In many other languages, however, the preposition determines the case of the noun phrase. In German, for example, in the equivalent in German, mit dem Mädchen, well, here, two requires the dative case. Dem is in the dative, and Mädchen is a possibility of expressing the dative in German. So the preposition, too, assigns case to its noun phrases. In English, however, we cannot see it. In German, if you're a learner of German, you have to learn it by heart. Each preposition is associated with a case that is assigned to an entire noun phrase. Let us now look at the verb. We know that the construction, speak very slowly to that girl, builds a constituent. Here are the reasons. We can use the entire construction, speak very slowly to that girl. And as a pro-form, so will I. So this boy will speak very slowly to that girl. So will I speak very slowly to that girl. Or we can apply the ellipsis test. Will you speak very slowly to that girl? Yes, I will speak very slowly to that girl. So we have reasons to believe that speak very slowly to that girl is another constituent. And since the verb is the head, speak is certainly the head, it determines, among others, its sister semantically. You can speak very slowly, but not very weakly. You can speak to that girl and not at that girl. Since this is the case, the verb is the head, and we have a new mother node, the verb phrase. Well, and the rest, the rest of the sentence, that's the combination of the two main branching nodes, noun phrase and verb phrase, with the auxiliary verb. And together they build the highest node in the tree, the sentence, the highest constituent, the sentence. There are several problems that are associated with this particular analysis or precisely with our suggestion of branching. The first problem is branching itself. We have constituents that are associated with two branches, so all those that are marked in red. And then we have the sentence that has three branches. Two are certainly necessary for mother nodes, for constituents. But is it necessary to have nodes that have three branches? Hmm, we will see. Another problem occurs when we combine this sentence with a complementizer. For example, if this boy will speak very slowly, where do we put if? Is it another branch that extends from the top level constituent S? Then we would have four branches. Or do we put it elsewhere, or do we have to change the system? Well, and then how should we label particular branching nodes? Here we had a branching node, which could also be called a verb phrase consisting of a verb and an adverb phrase where the verb is clearly the head. But what makes it different from the other verb phrase, which includes very slowly and to that girl? Hmm, these are just three problems. Here they are listed again. Branching, sentence initial elements, intermediate branching nodes. These are just three problems, and we will have to take care of them in another e-lecture. For the moment, it suffices. If you have understood the following, and that's my summary for this e-lecture. It should be clear now how you can argue for or against particular strings of words as constituents. You should know how we address the various parts of a syntactic tree, nodes, branches, mother nodes, daughter, sisters, and you should have a first idea about labeling constituents, all these phrases, noun phrases, adverb phrases, verb phrases, et cetera. So see you again in a follow-up e-lecture about constituent analysis, where we will deal with the problem of constituent hierarchies, and then we will also talk about intermediate phrasal categories. Until then, have a nice time. Bye-bye.