Four-vectors REFERENCE: Hartle Perhaps you've begun to think about vectors in (ct) four-dimensional space - after all, vectors are a major part of the language of physics. We will b call these vectors four-vectors to distinguish a + b them from the old three-dimensional vectors of Newtonian physics and Euclidean space. We will a also need to distinguish between the old three2a vectors and the new four vectors. We will -a reserve bold type from now on for four vectors, (x) and italicized normal type with a small arrow will be reserved for three-vectors. Thus V is a [Fig. 1] four-vector and V is a three-vector (and V is a scalar). Since it is hard to write in "bold" type, we will designate a hand-written four-vector by a squiggly underline, or an underline arrow. Thus V and V both represent the hand-written four-vector V. We define a four-vector as a directed line segment in four-dimensional Minkowski space just as we defined a three-vector as a directed line segment in Euclidean space. Four vectors can be graphically added and subtracted in the usual way, and they can be multiplied by a scalar to increase or decrease their length, or reverse their direction (see Fig. 1). Vectors can be of three types illustrated in Fig 2. Timelike vectors have a timelike separation between tail and tip, null vectors have null separation between tail and tip, and spacelike vectors have spacelike (ct) separation between tail and tip. The null vectors are so-named because, as at a et you may recall, the separation between two ey ex ay (x) points on a light line is zero (null). See "The new ax (y) spacetime geometry" if you [Fig. 3] have forgotten. (ct) Null Timelike Spacelike (x) (y) [Fig. 2] Just as we have defined the "basis three-vectors" i, j, and k of unit length in Euclidean space, we will define the basis four-vectors of unit length in Minkowski space: [Eq. 1] {et, ex, ey, ez} = {e0, e1, e2, e3} e. Note that basis vectors use subscripts, and components use superscripts. Any four-vector a can be specified by its components along the axes, a = (at, ax, ay, az) = (a0, a1, a2, a3) a, [Eq. 2] or in term of the basis four-vectors a = atet + axex + ayey + azez 3=0ae = ae. [Eq. 3] The last equality makes use of a special convention - namely that repeated upper and lower indices are understood to be summed. Greek indices (, , , , etc.) vary from 0 to 3 (through all four dimensions) and the Roman indices (i, j, k, etc.) vary from 1 to 3. Thus we can write the four-vector a as a = ae = a0e0 + a1e1 + a2e2 + a3e3 = atet + axex + ayey + azez. [Eq. 4] We can write the three-vector a as a = aiei = a1e1 + a2e2 + a3e3 = axex + ayey + azez. [Eq. 5] And we can write the four-vector a in terms of the three-vector a as a = a = (at, a ) = (at, ai ). Consider two events A and B located at coordinates (ctA,xA,yA,zA) and (ctB,xB,yB,zB), respectively (see Fig. 4). We then can define the displacement four-vector from A to B as x = (ctB - ctA, xB - xA, yB - yA, zB - zA), or [Eq. 7] x = xB - xA. [Eq. 6] (ct) B x A (x) [Fig. 4] Take note of the new use of the letter x. It is used as a generalized coordinate in Eq. 7. Thus xB2 yB, xA0 ctA, and so on. One must look at context to determine its usage. Eq. 7 is really four equations, since cycles through the values 0 through 3. Thus Eq. 7 is equivalent to x0 = xB0 - xA0, x1 = xB1 - xA1, x2 = xB2 - xA2, x3 = xB3 - xA3, or xt = xBt - xAt, xx = xBx - xAx, xy = xBy - xAy, xz = xBz - xAz, or ct = ctB - ctA, x = xB - xA, y = yB - yA, z = zB - zA. We call a dummy index. We could have expressed Eq. 7 as x = xB xA, or used any other Greek index. Indices that show up on both sides of an equality are cycled, producing one equation for each value for a total of four equations). Four-vectors can be transformed from one IRF to another moving at speed v relative to the positive x-axis via the Lorentz boost: ct' = (ct - xv/c), x' = (x - vt), y' = y, z' = z. For example, suppose we wish to transform the four vector given in (S) by a = (cat, ax, ay, az), into the corresponding four-vector in (S') by a' = (cat',ax',ay',az'). We simply match up the components with the correct coordinate. Thus cat ct, ax x, ay y, az z, and cat' ct', ax' x', ay' y', az' z', so that cat' = (cat - axv/c), ax' = (ax - vat), ay' = ay, az' = az. Finally, a' = (cat', ax', ay', az') = [(cat - axv/c) , (ax - vat), ay, az ].