You are on page 1of 4

Solutions to homework no.

1
E10 (a) It is easy to see that any two models with three points are isomorphic, just pick any bijective correspondence between its points, the correspondence between the lines follows from that uniquely. (b) See gure 1 for two non-isomorphic models with 4 points each (but with dierent numbers of lines).

Figure 1. Two non-isomorphic models for incidence geometry with 4 points each. (c) This is the isomorphism discussed in class in the context of constructing the projective plane associated to the ane plane given by the usual Euclidean plane. E12 (a) By incidence axiom 3 (I. A. 3), every model for incidence geometry needs to have at least 3 noncollinear points, say {A, B, C}. These pairwise determine 3 lines AB, AC and BC (see gure 2, rst picture). Since each of these lines contains only 2 points, we need to add at least 3 more points into our model, say {D, E, F } each lying on one of these lines (see gure 2, second picture). There is now a lack of lines in the new model. By I. A. 1 for each pair of points there needs to be a unique line passing through them. Consider rst the pairs {B, F }, {A, D} and {C, E}. For each of these 3 pairs, the corresponding lines AD, BF and CE cannot contain any other of the already existing points (see gure 2, third picture). For example if AD contained the point B then the lines AD and BD would intersect in 2 (or more) points contradicting proposition 2.1. Therefore in order to have all lines contain at least 3 points, we need an additional point, say G. To keep the number of
1

C F B A A

C D B E A

C D B E

C D B A E A

C D B E

Figure 2. Minimal model for incidence geometry where each line contains at least 3 points.

points at a minimum, we arrange G to lie on each of AD, BF

and CE (see gure 2, fourth picture). Finally, a check reveals that there are left over pairs of points with no lines passing through them, namely the pairs {E, F }, {E, D} and {D, F }. To remedy this, we add in another line into our model, one which contains all three of these points (see gure 2, fth picture). It is easy to verify that all the axioms of incidence geometry hold. There are 7 points and 7 lines in this model. Observe that this is the projective plane associated to the 4 point ane plane. (b) To construct the minimal geometry where the parallel postulate holds and where every line has at least three points, lets again start with the bare minimum of any model three points {A, B, C} and their associated lines AB, AC and BC (gure 3, rst picture). We now continue the construction by focusing rst on the parallel axiom. For the line AB and the point C there has to be a line parallel to AB and containing C. We add a point D to our model and let CD be that line. Likewise, let BD be the line parallel to AC and passing through B. In addition, let BD be the line through the points {B, D} (gure 3, second picture).

C H

G I

D F

Figure 3. Minimal model for incidence geometry where the parallel postulate holds and where each line contains at least 3 points. Neither of the lines so far contains 3 points. So we add the points {E, F, G, H, I} to ensure each line has 3 points (gure 3, third picture). We arrive at an interpretation which satises the parallel axiom and where each line has exactly 3 points. Unfortunately it fails the rst incidence axiom. For example the points G and H dont have any lines passing through them. To remedy this we add four more lines into the picture (gure 3, fourth picture). A case by case check shows that this is indeed a model of incidence geometry where the parallel axiom holds and where every line has precisely 3 points. Prop. 2.4 (1) Let P be any point and assume to the contrary of proposition 2.4 that every line contains P (RAA assumption). (2) The assumption of step (1) means that all lines are concurrent. This is a contradiction to proposition 2.2 and thus no such point P can exist. Prop. 2.5 (1) By incidence axiom 3 there exist 3 distinct non-collinear points {A, B, C}. (2) If P = A then the lines P B and P C both contain P . (3) If we had P B=P C then {P, B, C} (and hence also {A, B, C}) would be collinear which contradicts step (1). Thus we must have P B=P C and so we have two distinct lines containing P .

(4) If P = A consider the lines P A and P B. They both contain P . (5) If P A=P B consider in addition the line P C (which of course also contains P ). (6) If we had P A=P B and P A=P C then the points {P, A, B} are collinear as are the points {P, B, C} which implies that {A, B, C} are collinear contradicting step (1). (7) Thus either P A=P B or P A=P C. In either case we found two distinct lines containing P .

You might also like