General Relativity/BKL singularity

A BKL (Belinsky-Khalatnikov-Lifshitz) singularity[1] is a model of the dynamic evolution of the Universe near the initial singularity, described by an anisotropic, homogeneous, chaotic solution to Einstein's field equations of gravitation. According to this model, the Universe is oscillating (expanding and contracting) around a singular point (singularity) in which time and space become equal to zero. This singularity is physically real in the sense that it is a necessary property of the solution, and will appear also in the exact solution of those equations. The singularity is not artificially created by the assumptions and simplifications made by the other well-known special solutions such as the Friedmann-Lemaître-Robertson-Walker, quasi-isotropic, and Kasner solutions.

The Mixmaster universe is a solution to general relativity that exhibits properties similar to those discussed by BKL.

Existence of time singularity


The basis of modern cosmology are the special solutions of Einstein's field equations found by Alexander Friedmann in 1922 and 1924 that describe a completely homogeneous and isotropic Universe with any of two possible topologies corresponding to a space with a constant positive curvature ("closed model") or a constant negative curvature ("open model). The principal property of these solutions is their non-static nature. The concept of an inflating Universe that arises from Friedmann's solutions received a brilliant confirmation with the red-shift phenomenon discovered by E. Hubble and the present consensus is that the isotropic model, in general, gives an adequate description of the present state of the Universe.

At the same time, it is obvious that in the real world homogeneity is, at best, only an approximation. Even if one can speak about a homogeneous distribution of matter density at distances that are large compared to the intergalactic space, this homogeneity vanishes upon transition to smaller scales. On the other hand, the homogeneity assumption goes very far in a mathematical aspect. The high symmetry of the solution related to homogeneity can bring about specific properties that disappear when considering a more general case.

A related issue is how general is another important property of the isotropic model — the existence of a time singularity in the spacetime metric. In other words, the existence of such time singularity means finiteness of time. In the open model, there is one time singularity so time is limited from one end while in the closed model there are two singularities that limit time in both ends.

The adequacy of the isotropic model in describing the present state of the Universe by itself is not a reason to expect that it is so adequate in describing the early stages of Universe evolution. The problem initially addressed by the BKL paper[1] is whether the existence of such time singularity is a necessary property of relativistic cosmological models. There is the possibility that the singularity is generated by the simplifying assumptions, made when constructing these models. Independence of singularity on assumptions would mean that time singularity exists not only in the particular but also in the general solutions of the Einstein equations. A criterion for generality of solutions is the number of arbitrary space coordinate functions that they contain. These include only the "physically arbitrary" functions whose number cannot be reduced by any choice of reference frame. In the general solution, the number of such functions must be sufficient for arbitrary definition of initial conditions (distribution and movement of matter, distribution of gravitational field) in some moment of time chosen as initial. This number is four for vacuum and eight for a matter and/or radiation filled space.[2][3]

For a system of non-linear differential equations, such as the Einstein equations, general solution is not unambiguously defined. In principle, there may be multiple general integrals, and each of those may contain only a finite subset of all possible initial conditions. Each of those integrals may contain all required arbitrary functions which, however, may be subject to some conditions (e.g., some inequalities). Existence of a general solution with a singularity, therefore, does not preclude the existence also of other general solutions that do not contain a singularity. For example, there is no reason to doubt the existence of a general solution without singularity that describes an isolated body with a relatively small mass.

It is impossible to find a general integral for all space and for all time. However, this is not necessary for resolving the problem: it is sufficient to study the solution near the singularity. This would also resolve another aspect of the problem: the characteristics of spacetime metric evolution in the general solution when it reaches the physical singularity, understood as a point where matter density and invariants of the Riemann curvature tensor become infinite. The BKL paper[1] concerns only the cosmological aspect. This means, that the subject is a time singularity in the whole spacetime and not in some limited region as in a gravitational collapse of a finite body.

Previous work by the Landau-Lifshitz group[4][5][6] (reviewed in [2]) led to a conclusion that the general solution does not contain a physical singularity. This search for a broader class of solutions with singularity has been done, essentially, by a trial-and-error method, since a systemic approach to the study of the Einstein equations is lacking. A negative result, obtained in this way, is not convincing by itself; a solution with the necessary degree of generality would invalidate it, and at the same time would confirm any positive results related to the specific solution.

It is reasonable to suggest that if a singularity is present in the general solution, there must be some indications that are based only on the most general properties of the Einstein equations, although those indications by themselves might be insufficient for characterizing the singularity. At that time, the only known indication was related to the form of Einstein equations written in a synchronous reference frame, that is, in a frame in which the interval element is

  (eq. 1)

where the space distance element dl is separate from the time interval dt, and x0 = t is the proper time synchronized throughout the whole space.[7] The Einstein equation   written in synchronous frame gives a result in which the metric determinant g inevitably becomes zero in a finite time irrespective of any assumptions about matter distribution.[2][3]

This indication, however, was dropped after it became clear that it is linked with a specific geometric property of the synchronous frame: crossing of time line coordinates. This crossing takes place on some encircling hypersurfaces which are four-dimensional analogs of the caustic surfaces in geometrical optics; g becomes zero exactly at this crossing.[6] Therefore, although this singularity is general, it is fictitious, and not a physical one; it disappears when the reference frame is changed. This, apparently, stopped the incentive for further investigations.

However, the interest in this problem waxed again after Penrose published his theorems[8] that linked the existence of a singularity of unknown character with some very general assumptions that did not have anything in common with a choice of reference frame. Other similar theorems were found later on by Hawking[9][10] and Geroch[11] (see Penrose-Hawking singularity theorems). It became clear that the search for a general solution with singularity must continue.

Generalized Kasner solution


Further generalization of solutions depended on some solution classes found previously. The Friedmann solution, for example, is a special case of a solution class that contains three physically arbitrary coordinate functions.[2] In this class the space is anisotropic; however, its compression when approaching the singularity has "quasi-isotropic" character: the linear distances in all directions diminish as the same power of time. Like the fully homogeneous and isotropic case, this class of solutions exist only for a matter-filled space.

Much more general solutions are obtained by a generalization of an exact particular solution derived by Kasner[12] for a field in vacuum, in which the space is homogeneous and has Euclidean metric that depends on time according to the Kasner metric

  (eq. 2)

(see [13]). Here, p1, p2, p3 are any 3 numbers that are related by

  (eq. 3)

Because of these relationships, only 1 of the 3 numbers is independent. All 3 numbers are never the same; 2 numbers are the same only in the sets of values   and (0, 0, 1).[14] In all other cases the numbers are different, one number is negative and the other two are positive. If the numbers are arranged in increasing order, p1 < p2 < p3, they change in the ranges

  (eq. 4)

The numbers p1, p2, p3 can be written parametrically as

  (eq. 5)

All different values of p1, p2, p3 ordered as above are obtained by changing the value of the parameter u in the range u ≥ 1. The values u < 1 are brought into this range according to

  (eq. 6)
Figure 1

Figure 1 is a plot of p1, p2, p3 with an argument 1/u. The numbers p1(u) and p3(u) are monotonously increasing while p2(u) is monotonously decreasing function of the parameter u.

In the generalized solution, the form corresponding to (eq. 2) applies only to the asymptotic metric (the metric close to the singularity t = 0), respectively, to the major terms of its series expansion by powers of t. In the synchronous reference frame it is written in the form of (eq. 1) with a space distance element

  (eq. 7)

where   (eq. 8)

The three-dimensional vectors l, m, n define the directions at which space distance changes with time by the power laws (eq. 8). These vectors, as well as the numbers pl, pm, pn which, as before, are related by (eq. 3), are functions of the space coordinates. The powers pl, pm, pn are not arranged in increasing order, reserving the symbols p1, p2, p3 for the numbers in (eq. 5) that remain arranged in increasing order. The determinant of the metric of (eq. 7) is

  (eq. 9)

where v = l[mn]. It is convenient to introduce the following quantitities [15]

  (eq. 10)

The space metric in (eq. 7) is anisotropic because the powers of t in (eq. 8) cannot have the same values. On approaching the singularity at t = 0, the linear distances in each space element decrease in two directions and increase in the third direction. The volume of the element decreases in proportion to t.

The Einstein equations in vacuum in synchronous reference frame are[2][3]

  (eq. 11)
  (eq. 12)
  (eq. 13)

where   is the 3-dimensional tensor  , and Pαβ is the 3-dimensional Ricci tensor, which is expressed by the 3-dimensional metric tensor γαβ in the same way as Rik is expressed by gik; Pαβ contains only the space (but not the time) derivatives of γαβ.

The Kasner metric is introduced in the Einstein equations by substituting the respective metric tensor γαβ from (eq. 7) without defining a priori the dependence of a, b, c from t:


where the dot above a symbol designates differentiation with respect to time. The Einstein equation (eq. 11) takes the form

  (eq. 14)

All its terms are to a second order for the large (at t → 0) quantity 1/t. In the Einstein equations (eq. 12), terms of such order appear only from terms that are time-differentiated. If the components of Pαβ do not include terms of order higher than 2, then

  (eq. 15)

where indices l, m, n designate tensor components in the directions l, m, n.[2] These equations together with (eq. 14) give the expressions (eq. 8) with powers that satisfy (eq. 3).

However, the presence of 1 negative power among the 3 powers pl, pm, pn results in appearance of terms from Pαβ with an order greater than t−2. If the negative power is pl (pl = p1 < 0), then Pαβ contains the coordinate function λ and (eq. 12) become

  (eq. 16)

Here, the second terms are of order t−2(pm + pnpl) whereby pm + pnpl = 1 + 2 |pl| > 1.[16] To remove these terms and restore the metric (eq. 7), it is necessary to impose on the coordinate functions the condition λ = 0.

The remaining 3 Einstein equations (eq. 13) contain only first order time derivatives of the metric tensor. They give 3 time-independent relations that must be imposed as necessary conditions on the coordinate functions in (eq. 7). This, together with the condition λ = 0, makes 4 conditions. These conditions bind 10 different coordinate functions: 3 components of each of the vectors l, m, n, and one function in the powers of t (any one of the functions pl, pm, pn, which are bound by the conditions (eq. 3)). When calculating the number of physically arbitrary functions, it must be taken into account that the synchronous system used here allows time-independent arbitrary transformations of the 3 space coordinates. Therefore, the final solution contains overall 10 − 4 − 3 = 3 physically arbitrary functions which is 1 less than what is needed for the general solution in vacuum.

The degree of generality reached at this point is not lessened by introducing matter; matter is written into the metric (eq. 7) and contributes 4 new coordinate functions necessary to describe the initial distribution of its density and the 3 components of its velocity. This makes possible to determine matter evolution merely from the laws of its movement in an a priori given gravitational field. These movement laws are the hydrodynamic equations

  (eq. 17)
  (eq. 18)

where u i is the 4-dimensional velocity, ε and σ are the densities of energy and entropy of matter.[17] For the ultrarelativistic equation of state p = ε /3 the entropy σ ~ ε1/4. The major terms in (eq. 17) and (eq. 18) are those that contain time derivatives. From (eq. 17) and the space components of (eq. 18) one has


resulting in

  (eq. 19)

where 'const' are time-independent quantities. Additionally, from the identity uiui = 1 one has (because all covariant components of uα are to the same order)


where un is the velocity component along the direction of n that is connected with the highest (positive) power of t (supposing that pn = p3). From the above relations, it follows that

  (eq. 20)


  (eq. 21)

The above equations can be used to confirm that the components of the matter stress-energy-momentum tensor standing in the right hand side of the equations


are, indeed, to a lower order by 1/t than the major terms in their left hand sides. In the equations   the presence of matter results only in the change of relations imposed on their constituent coordinate functions.[2]

The fact that ε becomes infinite by the law (eq. 21) confirms that in the solution to (eq. 7) one deals with a physical singularity at any values of the powers p1, p2, p3 excepting only (0, 0, 1). For these last values, the singularity is non-physical and can be removed by a change of reference frame.

The fictional singularity corresponding to the powers (0, 0, 1) arises as a result of time line coordinates crossing over some 2-dimensional "focal surface". As pointed out in [2], a synchronous reference frame can always be chosen in such way that this inevitable time line crossing occurs exactly on such surface (instead of a 3-dimensional caustic surface). Therefore, a solution with such simultaneous for the whole space fictional singularity must exist with a full set of arbitrary functions needed for the general solution. Close to the point t = 0 it allows a regular expansion by whole powers of t.[18]

Oscillating mode towards the singularity


The four conditions that had to be imposed on the coordinate functions in the solution (eq. 7) are of different types: three conditions that arise from the equations   = 0 are "natural"; they are a consequence of the structure of Einstein equations. However, the additional condition λ = 0 that causes the loss of one derivative function, is of entirely different type.

The general solution by definition is completely stable; otherwise the Universe would not exist. Any perturbation is equivalent to a change in the initial conditions in some moment of time; since the general solution allows arbitrary initial conditions, the perturbation is not able to change its character. In other words, the existence of the limiting condition λ = 0 for the solution of (eq. 7) means instability caused by perturbations that break this condition. The action of such perturbation must bring the model to another mode which thereby will be most general. Such perturbation cannot be considered as small: a transition to a new mode exceeds the range of very small perturbations.

The analysis of the behavior of the model under perturbative action, performed by BKL, delineates a complex oscillatory mode on approaching the singularity.[1][19][20][21] They could not give all details of this mode in the broad frame of the general case. However, BKL explained the most important properties and character of the solution on specific models that allow far-reaching analytical study.

These models are based on a homogeneous space metric of a particular type. Supposing a homogeneity of space without any additional symmetry leaves a great freedom in choosing the metric. All possible homogeneous (but anisotropic) spaces are classified, according to Bianchi, in 9 classes.[22] BKL investigate only spaces of Bianchi Types VIII and IX.

If the metric has the form of (eq. 7), for each type of homogeneous spaces exists some functional relation between the reference vectors l, m, n and the space coordinates. The specific form of this relation is not important. The important fact is that for Type VIII and IX spaces, the quantities λ, μ, ν (eq. 10) are constants while all "mixed" products l rot m, l rot n, m rot l, etc. are zeros. For Type IX spaces, the quantities λ, μ, ν have the same sign and one can write λ = μ = ν = 1 (the simultaneous sign change of the 3 constants does not change anything). For Type VIII spaces, 2 constants have a sign that is opposite to the sign of the third constant; one can write, for example, λ = − 1, μ = ν = 1.[23]

The study of the effect of the perturbation on the "Kasner mode" is thus confined to a study on the effect of the λ-containing terms in the Einstein equations. Type VIII and IX spaces are the most suitable models exactly in this connection. Since all 3 quantities λ, μ, ν differ from zero, the condition λ = 0 does not hold irrespective of which direction l, m, n has negative power law time dependence.

The Einstein equations for the Type VIII and Type IX space models are[24]

  (eq. 22)
  (eq. 23)

(the remaining components  ,  ,  ,  ,  ,   are identically zeros). These equations contain only functions of time; this is a condition that has to be fulfilled in all homogeneous spaces. Here, the (eq. 22) and (eq. 23) are exact and their validity does not depend on how near one is to the singularity at t = 0.[25]

The time derivatives in (eq. 22) and (eq. 23) take a simpler form if а, b, с are substituted by their logarithms α, β, γ:

  (eq. 24)

substituting the variable t for τ according to:

  (eq. 25).


  (eq. 26)
  (eq. 27)

Adding together equations (eq. 26) and substituting in the left hand side the sum (α + β + γ)τ τ according to (eq. 27), one obtains an equation containing only first derivatives which is the first integral of the system (eq. 26):

  (eq. 28)

This equation plays the role of a binding condition imposed on the initial state of (eq. 26). The Kasner mode (eq. 8) is a solution of (eq. 26) when ignoring all terms in the right hand sides. But such situation cannot go on (at t → 0) indefinitely because among those terms there are always some that grow. Thus, if the negative power is in the function a(t) (pl = p1) then the perturbation of the Kasner mode will arise by the terms λ2a4; the rest of the terms will decrease with decreasing t. If only the growing terms are left in the right hand sides of (eq. 26), one obtains the system:

  (eq. 29)

(compare (eq. 16); below it is substituted λ2 = 1). The solution of these equations must describe the metric evolution from the initial state, in which it is described by (eq. 8) with a given set of powers (with pl < 0); let pl = р1, pm = р2, pn = р3 so that

  (eq. 30)


  (eq. 31)

where Λ is constant. Initial conditions for (eq. 29) are redefined as[26]

  (eq. 32)

Equations (eq. 29) are easily integrated; the solution that satisfies the condition (eq. 32) is

  (eq. 33)

where b0 and c0 are two more constants.

It can easily be seen that the asymptotic of functions (eq. 33) at t → 0 is (eq. 30). The asymptotic expressions of these functions and the function t(τ) at τ → −∞ is[27]


Expressing a, b, c as functions of t, one has

  (eq. 34)


  (eq. 35)


  (eq. 36)

The above shows that perturbation acts in such way that it changes one Kasner mode with another Kasner mode, and in this process the negative power of t flips from direction l to direction m: if before it was pl < 0, now it is p'm < 0. During this change the function a(t) passes through a maximum and b(t) passes through a minimum; b, which before was decreasing, now increases: a from increasing becomes decreasing; and the decreasing c(t) decreases further. The perturbation itself (λ2a in (eq. 29)), which before was increasing, now begins to decrease and die away. Further evolution similarly causes an increase in the perturbation from the terms with μ2 (instead of λ2) in (eq. 26), next change of the Kasner mode, and so on.

It is convenient to write the power substitution rule (eq. 35) with the help of the parametrization (eq. 5):

  (eq. 37)

The greater of the two positive powers remains positive.

BKL call this flip of negative power between directions a Kasner epoch. The key to understanding the character of metric evolution on approaching singularity is exactly this process of Kasner epoch alternation with flipping of powers pl, pm, pn by the rule (eq. 37).

The successive alternations (eq. 37) with flipping of the negative power p1 between directions l and m (Kasner epochs) continues by depletion of the whole part of the initial u until the moment at which u < 1. The value u < 1 transforms into u > 1 according to (eq. 6); in this moment the negative power is pl or pm while pn becomes the lesser of two positive numbers (pn = p2). The next series of Kasner epochs then flips the negative power between directions n and l or between n and m. At an arbitrary (irrational) initial value of u this process of alternation continues unlimited.[28]

In the exact solution of the Einstein equations, the powers pl, pm, pn lose their original, precise, sense. This circumstance introduces some "fuzziness" in the determination of these numbers (and together with them, to the parameter u) which, although small, makes meaningless the analysis of any definite (for example, rational) values of u. Therefore, only these laws that concern arbitrary irrational values of u have any particular meaning.

The larger periods in which the scales of space distances along two axes oscillate while distances along the third axis decrease monotonously, are called eras; volumes decrease by a law close to ~ t. On transition from one era to the next, the direction in which distances decrease monotonously, flips from one axis to another. The order of these transitions acquires the asymptotic character of a random process. The same random order is also characteristic for the alternation of the lengths of successive eras (by era length, BKL understand the number of Kasner epoch that an era contains, and not a time interval).

The era series become denser on approaching t = 0. However, the natural variable for describing the time course of this evolution is not the world time t but its logarithm, ln t, by which the whole process of reaching the singularity is extended to −∞.

According to (eq. 33), one of the functions a, b, c, that passes through a maximum during a transition between Kasner epochs, at the peak of its maximum is

  (eq. 38)

where it is supposed that amax is large compared to b0 and c0; in (eq. 38) u is the value of the parameter in the Kasner epoch before transition. It can be seen from here that the peaks of consecutive maxima during each era are gradually lowered. Indeed, in the next Kasner epoch this parameter has the value u' = u - 1, and Λ is substituted according to (eq. 36) with Λ' = Λ(1 − 2|p1(u)|). Therefore, the ratio of 2 consecutive maxima is


and finally

  (eq. 39)

The above are solutions to Einstein equations in vacuum. As for the pure Kasner mode, matter does not change the qualitative properties of this solution and can be written into it disregarding its reaction on the field.

However, if one does this for the model under discussion, understood as an exact solution of the Einstein equations, the resulting picture of matter evolution would not have a general character and would be specific for the high symmetry imminent to the present model. Mathematically, this specificity is related to the fact that for the homogeneous space geometry discussed here, the Ricci tensor components   are identically zeros and therefore the Einstein equations would not allow movement of matter (which gives non-zero stress energy-momentum tensor components  ).[29]

This difficulty is avoided if one includes in the model only the major terms of the limiting (at t → 0) metric and writes into it a matter with arbitrary initial distribution of densities and velocities. Then the course of evolution of matter is determined by its general laws of movement (eq. 17) and (eq. 18) that result in (eq. 21). During each Kasner epoch, density increases by the law

  (eq. 40)

where p3 is, as above, the greatest of the numbers p1, p2, p3. Matter density increases monotonously during all evolution towards the singularity.

To each era (s-th era) correspond a series of values of the parameter u starting from the greatest,  , and through the values   − 1,   − 2, ..., reaching to the smallest,   < 1. Then

  (eq. 41)

that is, k(s) = [ ] where the brackets mean the whole part of the value. The number k(s) is the era length, measured by the number of Kasner epochs that the era contains. For the next era

  (eq. 42)

In the limitless series of numbers u, composed by these rules, there are infinitesimally small (but never zero) values x(s) and correspondingly infinitely large lengths k(s).

Metric evolution


Very large u values correspond to Kasner powers

  (eq. 43)

which are close to the values (0, 0, 1). Two values that are close to zero, are also close to each other, and therefore the changes in two out of the three types of "perturbations" (the terms with λ, μ and ν in the right hand sides of (eq. 26)) are also very similar. If in the beginning of such long era these terms are very close in absolute values in the moment of transition between two Kasner epochs (or made artificially such by assigning initial conditions) then they will remain close during the greatest part of the length of the whole era. In this case (BKL call this the case of small oscillations), analysis based on the action of one type of perturbations becomes incorrect; one must take into account the simultaneous effect of two perturbation types.

Two perturbations


Consider a long era, during which 2 out of the 3 functions a, b, c (let them be a and b) undergo small oscillations while the third function (c) decreases monotonously. The latter function quickly becomes small; consider the solution just in the region where one can ignore c in comparison to a and b. The calculations are first done for the Type IX space model by substituting accordingly λ = μ = ν = 1.[20]

After ignoring function c, the first 2 equations (eq. 26) give

  (eq. 44)
  (eq. 45)

and as a third equation, (eq. 28) can be used, which takes the form

  (eq. 46)

The solution of (eq. 44) is written in the form


where α0, ξ0 are positive constants, and τ0 is the upper limit of the era for the variable τ. It is convenient to introduce further a new variable (instead of τ)

  (eq. 47)


  (eq. 48)

Equations (eq. 45) and (eq. 46) are transformed by introducing the variable χ = α − β:

  (eq. 49)
  (eq. 50)

Decrease of τ from τ0 to −∞ corresponds to a decrease of ξ from ξ0 to 0. The long era with close a and b (that is, with small χ), considered here, is obtained if ξ0 is a very large quantity. Indeed, at large ξ the solution of (eq. 49) in the first approximation by 1/ξ is

  (eq. 51)

where A is constant; the multiplier   makes χ a small quantity so it can be substituted in (eq. 49) by sh 2χ ≈ 2χ.[30]

From (eq. 50) one obtains


After determining α and β from (eq. 48) and (eq. 51) and expanding eα and eβ in series according to the above approximation, one obtains finally[31]:

  (eq. 52)
  (eq. 53)

The relation between the variable ξ and time t is obtained by integration of the definition dt = abc dτ which gives

  (eq. 54)

The constant c0 (the value of с at ξ = ξ0) should be now c0   α0·

Let us now consider the domain ξ   1. Here the major terms in the solution of (eq. 49) are:


where k is a constant in the range − 1 < k < 1; this condition ensures that the last term in (eq. 49) is small (sh 2χ contains ξ2k and ξ−2k). Then, after determining α, β, and t, one obtains

  (eq. 55)

This is again a Kasner mode with the negative t power coming into the function c(t).[32]

These results picture an evolution that is qualitatively similar to that, described above. During a long period of time that corresponds to a large decreasing ξ value, the two functions a and b oscillate, remaining close in magnitude  ; in the same time, both functions a and b slowly ( ) decrease. The period of oscillations is constant by the variable ξ : Δξ = 2π (or, which is the same, with a constant period by logarithmic time: Δ ln t = 2πΑ2). The third function, c, decreases monotonously by a law close to c = c0t/t0.

This evolution continues until ξ ~ 1 and formulas (eq. 52) and (eq. 53) are no longer applicable. Its time duration corresponds to change of t from t0 to the value t1, related to ξ0 according to

  (eq. 56)

The relationship between ξ and t during this time can be presented in the form

  (eq. 57)

After that, as seen from (eq. 55), the decreasing function c starts to increase while functions a and b start to decrease. This Kasner epoch continues until terms c2/a2b2 in (eq. 22) become ~ t2 and a next series of oscillations begins.

The law for density change during the long era under discussion is obtained by substitution of (eq. 52) in (eq. 20):

  (eq. 58)

When ξ changes from ξ0 to ξ ~ 1, the density increases   times.

It must be stressed that although the function c(t) changes by a law, close to c ~ t, the metric (eq. 52) does not correspond to a Kasner metric with powers (0, 0, 1). The latter corresponds to an exact solution (found by Taub[33]) which is allowed by eqs. 26-27 and in which

  (eq. 59)

where p, δ1, δ2 are constant. In the asymptotic region τ → −∞, one can obtain from here a = b = const, c = const.t after the substitution ерτ = t. In this metric, the singularity at t = 0 is non-physical.

Let us now describe the analogous study of the Type VIII model, substituting in eqs. 26-28 λ = −1, μ = ν = 1.[21]

If during the long era, the monotonically decreasing function is a, nothing changes in the foregoing analysis: ignoring a2 on the right side of equations (26) and (28), goes back to the same equations (49) and (50) (with altered notation). Some changes occur, however, if the monotonically decreasing function is b or c; let it be c.

As before, one has equation (49) with the same symbols, and, therefore, the former expressions (52) for the functions a(ξ) and b(ξ), but equation (50) is replaced by

  (eq. 60)

The major term at large ξ now becomes


so that

  (eq. 61)

The value of c as a function of time t is, as before c = c0t/t0 but the time dependence of ξ changes. The length of a long era depends on ξ0 according to

  (eq. 62)

On the other hand, the value ξ0 determines the number of oscillations of the functions a and b during an era (equal to ξ0/2π). Given the length of an era in logarithmic time (i.e., with given ratio t0/t1) the number of oscillations for Type VIII will be, generally speaking, less than for Type IX. For the period of oscillations one gets now Δ ln t = πξ/2; contrary to Type IX, the period is not constant throughout the long era, and slowly decreases along with ξ.

The small-time domain


As shown above, long eras violate the "regular" course of evolution; this fact makes it difficult to study the evolution of time intervals, encompassing several eras. It can shown, however, that such "abnormal" cases appear in the spontaneous evolution of the model to a singular point in the asymptotically small times t at sufficiently large distances from a start point with arbitrary initial conditions. Even in long eras both oscillatory functions during transitions between Kasner epochs remain so different that the transition occurs under the influence of only one perturbation. All results in this section relate equally to models of the types VIII and IX.[34]

During each Kasner epoch abc = Λt, i. e. α + β + γ = ln Λ + ln t. In transitions between epochs the constant ln Λ changes to the first order (cf. (eq. 36)). However, asymptotically to very large |ln t| values one can ignore not only these changes but also the constant ln Λ itself. In other words, this approximation corresponds to ignoring all values whose ratio to |ln t| converges to zero at t → 0. Then

  (eq. 63)

where Ω is the "logarithmic time"

  (eq. 64)

In this approximation, the process of epoch transitions can be regarded as a series of brief time flashes. The constant in the right hand side of condition (eq. 38) αmax = ½ ln (2|p1|Λ) that defines the periods of transition can also be ignored, i. e. this condition becomes α = 0 (or similar conditions for β or γ if the initial negative power is related to the functions b or c).[35] Thus, αmax, βmax, and γmax become zeros meaning that α, β, and γ will run only through negative values which are related in each moment by the relationship (eq. 64).

Figure 2

Considering such instant change of epochs, the transition periods are ignored as small in comparison to the epoch length; this condition is actually fulfilled.[36] Replacement of α, β, and γ maxima with zeros requires that quantities ln (|p1|Λ) be small in comparison with the amplitudes of oscillations of the respective functions. As mentioned above, during transitions between eras |p1| values can become very small while their magnitude and probability for occurrence are not related to the oscillation amplitudes in the respective moment. Therefore, in principle, it is possible to reach so small |p1| values that the above condition (zero maxima) is violated. Such drastic drop of αmax can lead to various special situations in which the transition between Kasner epochs by the rule (eq. 37) becomes incorrect (including the situations described above), see also [37]). These "dangerous" situations could break the laws used for the statistical analysis below. As mentioned, however, the probability for such deviations converges asymptotically to zero; this issue will be discussed below.

Consider an era that contains k Kasner epochs with a parameter u running through the values

  (eq. 65)

and let α and β are the oscillating functions during this era (Fig. 2).[38]

Initial moments of Kasner epochs with parameters un are Ωn. In each initial moment, one of the values α or β is zero, while the other has a minimum. Values α or β in consecutive minima, that is, in moments Ωn are

  (eq. 66)

(not distinguishing minima α and β). Values δn that measure those minima in respective Ωn units can run between 0 and 1. Function γ monotonously decreases during this era; according to (eq. 63) its value in moment Ωn is

  (eq. 67)

During the epoch starting at moment Ωn and ending at moment Ωn+1 one of the functions α or β increases from -δnΩn to zero while the other decreases from 0 to -δn+1Ωn+1 by linear laws, respectively:


resulting in the recurrent relationship

  (eq. 68)

and for the logarithmic epoch length

  (eq. 69)

where, for short, f(u) = 1 + u + u2. The sum of n epoch lengths is obtained by the formula

  (eq. 70)

It can be seen from (eq. 68) that |αn+1| > |αn|, i.e., the oscillation amplitudes of functions α and β increase during the whole era although the factors δn may be small. If the minimum at the beginning of an era is deep, the next minima will not become shallower; in other words, the residue |α — β| at the moment of transition between Kasner epochs remains large. This assertion does not depend upon era length k because transitions between epochs are determined by the common rule (eq. 37) also for long eras.

The last oscillation amplitude of functions α or β in a given era is related to the amplitude of the first oscillation by the relationship |αk-1| = |α0| (k + x) / (1 + x). Even at k 's as small as several units x can be ignored in comparison to k so that the increase of α and β oscillation amplitudes becomes proportional to the era length. For functions a = eα and b = eβ this means that if the amplitude of their oscillations in the beginning of an era was A0, at the end of this era the amplitude will become  .

The length of Kasner epochs (in logarithmic time) also increases inside a given era; it is easy to calculate from (eq. 69) that Δn+1 > Δn.[39] The total era length is

  (eq. 71)

(the term with 1/x arises from the last, k-th, epoch whose length is great at small x; cf. Fig. 2). Moment Ωn when the k-th epoch of a given era ends is at the same time moment Ω'0 of the beginning of the next era.

in the first Kasner epoch of the new era function γ is the first to rise from the minimal value γk = - Ωk (1 - δk) that it reached in the previous era; this value plays the role of a starting amplitude δ'0Ω'0 for the new series of oscillations. It is easily obtained that:

  (eq. 72)

It is obvious that δ'0Ω'0 > δ0Ω0. Even at not very great k the amplitude increase is very significant: function c = eγ begins to oscillate from amplitude  . The issue about the abovementioned "dangerous" cases of drastic lowering of the upper oscillation limit is left aside for now.

According to (eq. 40) the increase in matter density during the first (k - 1) epochs is given by the formula


For the last k epoch of a given era, it should be taken into account that at u = x < 1 the greatest power is p2(x) (not p3(x) ). Therefore, for the density increase over the whole era one obtains

  (eq. 73)

Therefore, even at not very great k values,  . During the next era (with a length k ' ) density will increase faster because of the increased starting amplitude A0':  , etc. These formulae illustrate the steep increase in matter density.

Statistical analysis near the singularity


The sequencing order of era lengths k(s), measured by the number of Kasner epochs contained in them, exhibits the character of a random process. The source of this stochasticity is the rule (eq. 41-42) according to which the transition from one era to the next is determined from an infinite numerical sequence of u values.

In the statistical description of this sequence, instead of a fixed initial value umax = k(0) + x(0), BKL consider values of x(0) that are distributed in the interval from 0 to 1 by some probabilistic distributional law. Then the values of x(s) that finish each (s-th) number series will also be distributed according to some laws. It can be shown [1] that with growing s these distributions converge to a definite static (s-independent) distribution of probabilities w(x) in which the initial conditions are completely "forgotten":

  (eq. 74)

This allows to find the distribution of probabilities for length k:

  (eq. 75)

The above formulae are the basis on which the statistical properties of the model evolution are studied.[34]

This study is complicated by the slow decrease of the distribution function (eq. 75) at large k:

  (eq. 76)

The mean value  , calculated from this distribution, diverges logarithmically. For a sequence, cut off at a very large but still finite number N, one has  . The usefulness of the mean in this case is very limited because of its instability: because of the slow decrease of W(k), fluctuations in k diverge faster than its mean. A more adequate characteristic of this sequence is the probability that a randomly chosen number from it belongs to a series of length K where K is large. This probability is lnK/lnN. It is small if  . In this respect one can say that a randomly chosen number from the given sequence belongs to the long series with a high probability.

The recurrent formulae defining transitions between eras are re-written and detailed below. Index s numbers the successive eras (not the Kasner epochs in a given era!), beginning from some era (s = 0) defined as initial. Ω(s) and ε(s) are, respectively, the initial moment and initial matter density in the s-th era; δsΩs is the initial oscillation amplitude of that pair of functions α, β, γ, which oscillates in the given era: k(s) is the length of s-th era, and x(s) determines the length of the next era according to k(s+1) = [1/x(s)]. According to (eq. 71-73)

  (eq. 77)
  (eq. 78)
  (eq. 79)

s is introduced in (eq. 77) to be used further on).

The values of δ(s) (ranging from 0 to 1) have their own static statistical distribution. It satisfies an integral equation expressing the fact that δ(s) and δ(s+1) which are related through (eq. 78) have an identical distribution; this equation can be solved numerically (cf. [34]). Since (eq. 78) does not contain a singularity, the distribution is perfectly stable; the mean values of δ or its powers calculated through it are definite finite numbers. In particular, the mean value of δ is  

The statistical relation between large time intervals Ω and the number of eras s contained in them is found by repeated application of (eq. 77):

  (eq. 80)

Direct averaging of this equation, however, does not make sense: because of the slow decrease of function W(k) mean values of exp(ξs) are unstable in the above sense. This instability is removed by taking logarithm: the "double-logarithmic" time interval

  (eq. 81)

is expressed by the sum of values ξp which have a stable statistical distribution. The mean values of ξs and their powers (calculated from the distributions of values x, k and δ) are finite; numeric calculation gives  

Averaging (eq. 81) at a given s obtains

  (eq. 82)

which determines the mean double-logarithmic time interval containing s successive eras.

In order to calculate the mean square of fluctuations of this value one writes


In the last equation, it is taken into account that in the static limit the statistical correlation between ξ(s) and ξ′(s) depends only on the difference | s - s′ |. Due to the existing recurrent relationship between x(s), k(s), δ(s) and x(s+1), k(s+1), δ(s+1) this correlation is, strictly speaking, different from zero. It, however, quickly decreases with increasing | s - s′ | and numeric calculation shows that even at | s - s′ | = 1,   = − 0.4. Leaving the first two terms in the sum by p, one obtains

  (eq. 83)

At s → ∞ the relative fluctuation (i.e., the ratio between the mean squared fluctuations (eq. 83) and the mean value (eq. 82)), therefore, approaches zero as s-1/2. In other words, the statistical relationship (eq. 82) at large s becomes close to certainty. This is a corollary that according to (eq. 81) τs can be presented as a sum of a large number of quasi-independent additives (i.e., it has the same origin as the certainty of the values of additive thermodynamic properties of macroscopic bodies). Therefore, the probabilities of various τs values (at given s) have a Gaussian distribution:

  (eq. 84)

Certainty of relationship (eq. 82) allows its reversal, i.e., express it as a dependence of the mean number of eras   contained in a given interval of double-logarithmic time τ:

  (eq. 85)

The respective statistical distribution is given by the same Gaussian distribution in which the random variable is now sτ at a given τ:

  (eq. 86)

Respective to matter density, (eq. 79) can be re-written with account of (eq. 80) in the form


and then, for the complete energy change during s eras,

  (eq. 87)

The term with the sum by p gives the main contribution to this expression because it contains an exponent with a large power. Leaving only this term and averaging (eq. 87), one gets in its right hand side the expression   which coincides with (eq. 82); all other terms in the sum (also terms with ηs in their powers) lead only to corrections of a relative order 1/s. Therefore

  (eq. 88)

Thanks to the above established almost certain character of the relation between τs and s (eq. 88) can be written as


which determines the value of the double logarithm of density increase averaged by given double-logarithmic time intervals τ or by a given number of eras s.

These stable statistical relationships exist specifically for double-logarithmic time intervals and for the density increase. For other characteristics, e.g., ln (ε(s)(0)) the relative fluctuation increase by a power law with the increase of the averaging range thereby devoiding the term mean value of its sense of stability.

As shown below, in the limiting asymptotic case the abovementioned "dangerous" cases that disturb the regular course of evolution expressed by the recurrent relationships (eq. 77-79), do not occur in reality.

Dangerous are cases when at the end of an era the value of the parameter u = x (and with it also |p1| ≈ x). A criterion for selection of such cases is the inequality

  (eq. 89)

where | α(s) | is the initial minima depth of the functions that oscillate in era s (it would have been better to take the final amplitude but that would only strengthen the selection criterion).

The value of x(0) in the first era is determined by the initial conditions. Dangerous are values in the interval δx(0) ~ exp ( − | α(0) | ), and also in intervals that could result in dangerous cases in the next eras. In order that x(s) comes into the dangerous interval δx(s) ~ exp ( − | α(s) | ), the initial value x(0) should lie into an interval of a width δx(0) ~ δx(s) / k(1)^2 ... k(s)^2.[40] Therefore, from a unit interval of all possible values of x(0), dangerous cases will appear in parts λ of this interval:

  (eq. 90)

(the inner sum is taken by all values k(1), k(2), ... , k(s) from 1 to ∞). It is easy to show that this series converges to the value λ   1 whose order of magnitude is determined by the first term in (eq. 90). This can be shown by a strong majoration of the series for which one substitutes | α(s) | = (s+1) | α(0) |, regardless of the lengths of eras k(1), k(2), ... (In fact | α(s) | increase much faster; even in the most unfavorable case k(1) = k(2) = ... = 1 values of | α(s) | increase as qs | α(0) | with q > 1.) Noting that


one obtains


If the initial value of x(0) lies outside the dangerous region λ there will be no dangerous cases. If it lies inside this region dangerous cases occur but upon their completion the model resumes a "regular" evolution with a new initial value which only occasionally (with a probability λ) may come into the dangerous interval. Repeated dangerous cases occur with probabilities λ2, λ3, ... , asymptopically converging to zero.

General solution with small oscillations


In the above models, metric evolution near the singularity is studied on the example of homogeneous space metrics. It is clear from the characteristic of this evolution that the analytic construction of the general solution for a singularity of such type should be made separately for each of the basic evolution components: for the Kasner epochs, for the process of transitions between epochs caused by "perturbations", for long eras with two perturbations acting simultaneously. During a Kasner epoch (i.e. at small perturbations), the metric is given by (eq. 7) without the condition λ = 0.

BKL further developed a matter distribution-independent model (homogeneous or non-homogeneous) for long era with small oscillations. The time dependence of this solution turns out to be very similar to that in the particular case of homogeneous models; the latter can be obtained from the distribution-independent model by a special choice of the arbitrary functions contained in it. [41]

It is convenient, however, to construct the general solution in a system of coordinates somewhat different from synchronous reference frame: g = 0 as in the synchronous frame but instead of g00 = 1 it is now g00 = − g33. Defining again the space metric tensor γαβ = − gαβ one has, therefore

  (eq. 91)

The special space coordinate is written as x3 = z and the time coordinate is written as x0 = ξ (as different from proper time t); it will be shown that ξ corresponds to the same variable defined in homogeneous models. Differentiation by ξ and z is designated, respectively, by dot and prime. Latin indices a, b, c take values 1, 2, corresponding to space coordinates x1, x2 which will be also written as x, y. Therefore, the metric is

  (eq. 92)

The required solution should satisfy the inequalities

  (eq. 93)
  (eq. 94)

(these conditions specify that one of the functions a2, b2, c2 is small compared to the other two which was also the case with homogeneous models).

Inequality (eq. 94) means that components γa3 are small in the sense that at any ratio of the shifts dxa and dz, terms with products dxadz can be omitted in the square of the spatial length element dl2. Therefore, the first approximation to a solution is a metric (eq. 92) with γa3 = 0:[42]

  (eq. 95)

One can be easily convinced by calculating the Ricci tensor components  ,  ,  ,   using metric (eq. 95) and the condition (eq. 93) that all terms containing derivatives by coordinates xa are small compared to terms with derivatives by ξ and z (their ratio is ~ γ33 / γab). In other words, to obtain the equations of the main approximation, γ33 and γab in (eq. 95) should be differentiated as if they do not depend on xa. Designating

  (eq. 96)

one obtains the following equations:[43]

  (eq. 97)
  (eq. 98)
  (eq. 99)

Index raising and lowering is done here with the help of γab. The quantities   and λ are the contractions   and   whereby

  (eq. 100)

As to the Ricci tensor components  ,  , by this calculation they are identically zero. In the next approximation (i.e., with account to small γa3 and derivatives by x, y), they determine the quantities γa3 by already known γ33 and γab.

Contraction of (eq. 97) gives  , and, hence,

  (eq. 101)

Different cases are possible depending on the G variable. In the above case g00 = γ33   γab and  . The case N > 0 (quantity N is time-like) leads to time singularities of interest. Substituting in (eq. 101) f1 = 1/2 ( ξ + z ) sin y, f2 = 1/2 ( ξ - z ) sin y results in G of type

  (eq. 102)

This choice does not diminish the generality of conclusions; it can be shown that generality is possible (in the first approximation) just on account of the remaining permissible transformations of variables. At N < 0 (quantity N is space-like) one can substitute G = z which generalizes the well known Einstein-Rosen metric.[44] At N = 0 one arrives at the Robinson-Bondi wave metric that depends only on ξ + z or only on ξ - z (cf. [45]). The factor sin y in (eq. 102) is put for convenient comparison with homogeneous models. Taking into account (eq. 102), equations 97-99 become

  (eq. 103)
  (eq. 104)
  (eq. 105)

The principal equations are (eq. 103) defining the γab components; then, function ψ is found by a simple integration of (eq. 104-105).

The variable ξ runs through the values from 0 to ∞. The solution of (eq. 103) is considered at two boundaries, ξ   1 and   1. At large ξ values, one can look for a solution that takes the form of a 1 / √ξ decomposition:

  (eq. 106)


  (eq. 107)

(equation 107 needs condition 102 to be true). Substituting (eq. 103) in (eq. 106), one obtains in the first order

  (eq. 108)

where quantities aac constitute a matrix that is inverse to matrix aac. The solution of (eq. 108) has the form

  (eq. 109)
  (eq. 110)

where la, ma, ρ, are arbitrary functions of coordinates x, y bound by condition (eq. 110) derived from (eq. 107).

To find higher terms of this decomposition, it is convenient to write the matrix of required quantities γab in the form

  (eq. 111)
  (eq. 112)

where the symbol ~ means matrix transposition. Matrix H is symmetric and its trace is zero. Presentation (eq. 111) ensures symmetry of γab and fulfillment of condition (eq. 102). If exp H is substituted with 1, one obtains from (eq. 111) γab = ξaab with aab from (eq. 109). In other words, the first term of γab decomposition corresponds to H = 0; higher terms are obtained by powers decomposition of matrix H whose components are considered small.

The independent components of matrix H are written as σ and φ so that

  (eq. 113)

Substituting (eq. 111) in (eq. 103) and leaving only terms linear by H, one derives for σ and φ

  (eq. 114)

If one tries to find a solution to these equations as Fourier series by the z coordinate, then for the series coefficients, as functions of ξ, one obtains Bessel equations. The major asymptotic terms of the solution at large ξ are[46]

  (eq. 115)

Coefficients A and B are arbitrary complex functions of coordinates x, y and satisfy the necessary conditions for real σ and φ; the base frequency ω is an arbitrary real function of x, y. Now from (eqs. 104-105) it is easy to obtain the first term of the function ψ:

  (eq. 116)

(this term vanishes if ρ = 0; in this case the major term is the one linear for ξ from the decomposition: ψ = ξq (x, y) where q is a positive function[33]).

Therefore, at large ξ values, the components of the metric tensor γab oscillate upon decreasing ξ on the background of a slow decrease caused by the decreasing ξ factor in (eq. 111). The component γ33 = eψ decreases quickly by a law close to exp (ρ2ξ2); this makes it possible for condition (eq. 93).[47]

Next BKL consider the case ξ   1. The first approximation to a solution of (eq. 103) is found by the assumption (confirmed by the result) that in these equations terms with derivatives by coordinates can be left out:

  (eq. 117)

This equation together with the condition (eq. 102) gives

  (eq. 118)

where λa, μa, s1, s2 are arbitrary functions of all 3 coordinates x, y, z, which are related with other conditions

  (eq. 119)

Equations 104-105 give now

  (eq. 120)

The derivatives  , calculated by (eq. 118), contain terms ~ ξ4s1-2 and ~ ξ4s2-2 while terms left in (eq. 117) are ~ ξ-2. Therefore, application of (eq. 103) instead of (eq. 117) is permitted on conditions s1 > 0, s2 > 0; hence 1 -   > 0.

Thus, at small ξ oscillations of functions γab cease while function γ33 begins to increase at decreasing ξ. This is a Kasner mode and when γ33 is compared to γab, the above approximation is not applicable.

In order to check the compatibility of this analysis, BKL studied the equations   = 0,   = 0, and, calculating from them the components γa3, confirmed that the inequality (eq. 94) takes place. This study[41] showed that in both asymptotic regions the components γa3 were ~ γ33. Therefore, correctness of inequality (eq. 93) immediately implies correctness of inequality (eq. 94).

This solution contains, as it should be for the general case of a field in vacuum, four arbitrary functions of the three space coordinates x, y, z. In the region ξ   1 these functions are, e.g., λ1, λ2, μ1, s1. In the region ξ   1 the four functions are defined by the Fourier series by coordinate z from (eq. 115) with coefficients that are functions of x, y; although Fourier series decomposition (or integral?) characterizes a special class of functions, this class is large enough to encompass any finite subset of the set of all possible initial conditions.

The solution contains also a number of other arbitrary functions of the coordinates x, y. Such two-dimensional arbitrary functions appear, generally speaking, because the relationships between three-dimensional functions in the solutions of the Einstein equations are differential (and not algebraic), leaving aside the deeper problem about the geometric meaning of these functions. BKL did not calculate the number of independent two-dimensional functions because in this case it is hard to make unambiguous conclusions since the three-dimensional functions are defined by a set of two-dimensional functions (cf. [41] for more details).[48]

Finally, BKL go on to show that the general solution contains the particular solution obtained above for homogeneous models.

Substituting the basis vectors for Bianchi Type IX homogeneous space in (eq. 7) the space-time metric of this model takes the form

  (eq. 121)

When c2   a2, b2, one can ignore c2 everywhere except in the term c2 dz2. To move from the synchronous frame used in (eq. 121) to a frame with conditions (eq. 91), the transformation dt = c dξ/2 and substitution zz/2 are done. Assuming also that χ ≡ ln (a/b)   1, one obtains from (eq. 121) in the first approximation:

  (eq. 122)

Similarly, with the basis vectors of Bianchi Type VIII homogeneous space, one obtains

  (eq. 123)

According to the analysis of homogeneous spaces above, in both cases ab = ξ (simplifying   = ξ0) and χ is from (eq. 51); function c (ξ) is given by formulae (eq. 53) and (eq. 61), respectively, for models of Types IX and VIII.

Identical metric for Type VIII is obtained from (eq. 112, 115, 116) choosing two-dimensional vectors la and ma in the form

  (eq. 124)

and substituting

  (eq. 125)

To obtain the metric for Type IX, one should substitute

  (eq. 126)

(for calculation of c (ξ) the approximation in (eq. 116) is not sufficient and the term in ψ linear by ξ is calculated[33])

This analysis was done for empty space. Including matter does not make the solution less general and does not change its qualitative characteristics.[33][41]



BKL describe singularities in the cosmologic solution of Einstein equations that have a complicated oscillatory character. Although this singularity was studied primarily on special homogeneous models, there are convincing reasons to assume that singularities in the general solution of Einstein equations have the same characteristics; this circumstance makes the BKL model important for cosmology.

A basis for such statement is the fact that the oscillatory mode in the approach to singularity is caused by the single perturbation that also causes instability in the generalized Kasner solution. A confirmation of the generality of the model is the analytic construction for long era with small oscillations. Although this latter behavior is not a necessary element of metric evolution close to the singularity, it has all principal qualitative properties: metric oscillation in two spacial dimensions and monotonous change in the third dimension with a certain perturbation of this mode at the end of some time interval. However, the transitions between Kasner epochs in the general case of non-homogeneous spacial metric have not been elucidated in details.

The problem connected with the possible limitations upon space geometry caused by the singularity was left aside for further study. It is clear from the outset, however, that the original BKL model is applicable to both finite or infinite space; this is evidenced by the existence of oscillatory singularity models for both closed and open spacetimes.

The oscillatory mode of the approach to singularity gives a new aspect to the term 'finiteness of time'. Between any finite moment of the world time t and the moment t = 0 there is an infinite number of oscillations. In this sense, the process acquires an infinite character. Instead of time t, a more adequate variable for its description is ln t by which the process is extended to − ∞.

BKL consider metric evolution in the direction of decreasing time. The Einstein equations are symmetric in respect to the time sign so that a metric evolution in the direction of increasing time is equally possible. However, these two cases are fundamentally different because past and future are not equivalent in the physical sense. Future singularity can be physically meaningful only if it is possible at arbitrary initial conditions existing in a previous moment. Matter distribution and fields in some moment in the evolution of Universe do not necessarily correspond to the specific conditions required for the existence of a given special solution to the Einstein equations.

The choice of solutions corresponding to the real world is related to profound physical requirements which is impossible to find using only the existing relativity theory and which can be found as a result of future synthesis of physical theories. Thus, it may turn out that this choice singles out some special (e.g., isotropic) type of singularity. Nevertheless, it is more natural to assume that because of its general character, the oscillatory mode should be the main characteristic of the initial evolutionary stages.

In this respect, of considerable interest is the property of the model, shown by Misner[49], related to propagation of light signals. In the isotropic model, a "light horizont" exists, meaning that for each moment of time, there is some longest distance, at which exchange of light signals and, thus, a causal connection, is impossible: the signal cannot reach such distances for the time since the singularity t = 0.

Signal propagation is determined by the equation ds = 0. In the isotropic model near the singularity t = 0 the interval element is ds2 = dt2 — 2t  , where   is a time-independent spatial differential form [50]. Substituting t = η2/2 yields

  (eq. 127)

The "distance" Δ  reached by the signal is

  (eq. 128)

Since η, like t, runs through values starting from 0, up to the "moment" η signals can propagate only at the distance Δ  ≤ η which fixes the farthest distance to the horizon.

The existence of a light horizon in the isotropic model poses a problem in the understanding of the origin of the presently observed isotropy in the relic radiation. According to the isotropic model, the observed isotropy means isotropic properties of radiation that comes to the observer from such regions of space that can not be causally connected with each other. The situation in the oscillatory evolution model near the singularity can be different.

For example, in the homogeneous model for Type IX space, a signal is propagated in a direction in which for a long era, scales change by a law close to ~ t. The square of the distance element in this direction is dl2 = t2 , and the respective element of the four-dimensional interval is ds2 = dt2t2 . The substitution t = еη puts this in the form

  (eq. 129)

and for the signal propagation one has equation of the type (eq. 128) again. The important difference is that the variable η runs now through values starting from − ∞ (if metric (eq. 129) is valid for all t starting from t = 0).

Therefore, for each given "moment" η are found intermediate intervals Δη sufficient for the signal to cover each finite distance.

In this way, during a long era a light horizon is opened in a given space direction. Although the duration of each long era is still finite, during the course of the world evolution eras change an infinite number of times in different space directions. This circumstance makes one expect that in this model a causal connection between events in the whole space is possible. Because of this property, Misner named this model "mixmaster universe" by a brand name of a dough-blending machine.

As time passes and one goes away from the singularity, the effect of matter on metric evolution, which was insignificant at the early stages of evolution, gradually increases and eventually becomes dominant. It can be expected that this effect will lead to a gradual "isotropisation" of space as a result of which its characteristics come closer to the Friedman model which adequately describes the present state of the Universe.

Finally, BKL pose the problem about the feasibility of considering a "singular state" of a world with infinitely dense matter on the basis of the existing relativity theory. The physical application of the Einstein equations in their present form in these conditions can be made clear only in the process of a future synthesis of physical theories and in this sense the problem can not be solved at present.

It is important that the gravitational theory itself does not lose its logical cohesion (i.e., does not lead to internal controversies) at whatever matter densities. In other words, this theory is not limited by the conditions that it imposes, which could make logically inadmissible and controversial its application at very large densities; limitations could, in principle, appear only as a result of factors that are "external" to the gravitational theory. This circumstance makes the study of singularities in cosmological models formally acceptable and necessary in the frame of existing theory.


  1. a b c d e Belinsky, Vladimir A.; Khalatnikov, Isaak M.; Lifshitz, Evgeny M. (1970), "Колебательный режим приближения к особой точке в релятивистской космологии", Uspekhi Fizicheskikh Nauk (Успехи Физических Наук), 102(3) (11): 463–500 {{citation}}: Cite has empty unknown parameter: |1= (help); English translation in Belinskii, V.A. (1970). "Oscillatory Approach to a Singular Point in the Relativistic Cosmology". Advances in Physics. 19: 525–573. doi:10.1080/00018737000101171.
  2. a b c d e f g h Lifshitz, Khalatnikov (1963)
  3. a b c Landau, Lifshitz (1988), Section 97, Synchronous reference frame
  4. Lifshitz, Evgeny M. (1960). Zhurnal' Eksperimental'noy i Teoreticheskoy Fiziki. 39: 149. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  5. Lifshitz, Evgeny M. (1960). Zhurnal' Eksperimental'noy i Teoreticheskoy Fiziki. 39: 800. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  6. a b Lifshitz, Evgeny M. (1961). Zhurnal' Eksperimental'noy i Teoreticheskoy Fiziki. 40: 1847. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help); Phys. Rev. Letts., 6, 311 (1961)
  7. The convention used by BKL is the same as in the Landau, Lifshitz (1988) book. The Latin indices run through the values 0, 1, 2, 3; Greek indices run through the space values 1, 2, 3. The metric gik has the signature (+ − − −); γαβ = − gαβ is the 3-dimensional space metric tensor. BKL use a system of units, in which the speed of light and the Einstein gravitational constant are equal to 1.
  8. Penrose, Roger (1965). Physical Review Letters. 14: 57. {{cite journal}}: Missing or empty |title= (help)
  9. Hawking, Stephen W. (1965). Physical Review Letters. 15: 689. {{cite journal}}: Missing or empty |title= (help)
  10. Hawking, Stephen W.; Ellis, G.F.R. (1968). Astrophysical Journal. 152: 25. {{cite journal}}: Missing or empty |title= (help)
  11. Geroch, Robert P. (1966). Physical Review Letters. 17: 445. {{cite journal}}: Missing or empty |title= (help)
  12. Kasner, Edward (1921). American Journal of Mathematics. 43. {{cite journal}}: Missing or empty |title= (help)
  13. Landau, Lifshitz (1988), Section 117, Flat anisotropic model
  14. When (p1, p2, p3) = (0, 0, 1) the spacetime metric (eq. 1) with dl2 from (eq. 2) transforms to Galilean metric with the substitution t sh z = ζ, t ch z = τ, that is, the singularity is fictional and the spacetime is flat.
  15. Here and below all symbols for vector operations (vector products, the operations rot, grad, etc.) should be understood in a very formal way as operations over the covariant components of the vectors l, m, n such that are performed in Cartesian coordinates x1, x2, x3.
  16. Excepting the case (p1, p2, p3) = (0, 0, 1), in which the metric singularity is fictitious.
  17. cf. Misner, Charles W. (1973). Gravitation. San Francisco: W.H. Freeman and Company. p. 564. ISBN 0-7167-0334-3. {{cite book}}: Unknown parameter |coauthors= ignored (|author= suggested) (help)
  18. For an analysis of this case, see Belinsky, Vladimir A. (1965). Zhurnal' Eksperimental'noy i Teoreticheskoy Fiziki. 49: 1000. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  19. Khalatnikov, I.M. (1970). Physical Review Letters. 24: 76. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  20. a b Belinsky, Vladimir A. (1969). Zhurnal' Eksperimental'noy i Teoreticheskoy Fiziki. 56: 1700. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  21. a b Lifshitz, Evgeny M. (1970). Pis'ma Zhurnalya Eksperimental'noy i Teoreticheskoy Fiziki. 11: 200. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  22. Belinsky, Khalatnikov, Lifshitz (1970), Appendix C
  23. The constants λ, μ, ν are the so-called structural constants of the space movement group.
  24. Lifshitz, Khalatnikov (1963), Appendix C
  25. In their exact form, the Einstein equations for homogeneous space contain, in general, 6 different functions of time γab(t) in the metric. The fact that in the present case a consistent system of exact equations is obtained for the metric which contains only 3 functions of time (γ11 = а2, γ22 = b2, γ33 = c2) is related to a symmetry that leads to the disappearance of 6 Ricci tensor components.
  26. It should be reminded that BKL model the evolution in the direction t → 0; therefore, the "initial" conditions exist at later rather than at earlier times.
  27. The asymptotic values of ατ, βτ, γτ at τ → −∞ can be found without fully solving (eq. 29). It suffices to note that the first of these equations has a form of a "particle" moving in one dimension in the field of an exponential potential wall with α playing the role of a constant. In this analogy, the Kasner mode refers to a free movement with constant velocity ατ = Λp1. After reflection from the wall, the particle moves freely with velocity ατ = −Λp1. Also noting that from (eq. 29) ατ + βτ = const, and ατ + γτ = const, one can see that βτ and γτ take the values βτ = Λ(p2 − 2p1), γτ = Λ(p3 − 2p1).
  28. Introduction of non-diagonal components of γab(t) imparts some new features to the BKL model: rotations of axes corresponding to the Kasner epoch powers; this problem is studied in Belinsky, Vladimir A. (1971). Zhurnal' Eksperimental'noy i Teoreticheskoy Fiziki. 60 (3). {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  29. In other words, the synchronous frame must also be co-moving with respect to matter. If one substitutes in (eq. 19) uα = 0, u0 = 1, it becomes ε ~ (abc)-4/3 ~ t-4/3.
  30. The constant in the sine argument, of course, is not necessarily the same as ξ0 in (eq. 47) and (eq. 48); however, making them the same does not change in any way the solution character.
  31. In a more precise calculation, a slowly changing logarithmic term appears in the sine argument, and a multiplier appears in front the exponent in the expression for с(ξ), see Belinsky, Khalatnikov, Lifshitz (1970), Appendix B.
  32. If in (eq. 49), one substitutes sh 2χ with 2χ and solves it for all values of ξ, one obtains χ = c1J0(ξ) + c2N0(ξ) where J0, N0 are Bessel functions of the I and II kind. This solution interpolates between the two limiting cases and allows to relate by an order of magnitude the constant parameters in (eq. 52) and (eq. 55).
  33. a b c d Belinsky, V.A. (1969). Zhurnal' Eksperimental'noy i Teoreticheskoy Fiziki. 57: 2163. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  34. a b c Lifshitz, E.M. (1970). Zhurnal' Eksperimental'noy i Teoreticheskoy Fiziki. 59: 322. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  35. This means ignoring the effect of gradually decreasing maxima of the oscillating functions during an era described by the formula (eq. 39)
  36. Also note, that epoch lengths are large in comparison with transitions between epochs. According to (eq. 32), transitions are large with small |p1| (i. e. large u) and are ~ 1/|p1| ~ u. But even in this case Δn ~ unn|   un
  37. Doroshkevich, A.G. (1970). Astronomicheskiy Zhurnal. 47 (5). {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  38. Fixing the limits of the era according to eq. 64 is meaningful because in such case the era contains all epochs in which the third function, γ(t) decreases monotonously. If the era is defined by the sequence of u values from k + x to 1 + x, then the monotonous decrease of γ(t) will continue during the first epoch of the next era.
  39. Epoch lengths are great in comparison to transitions between epochs. According to (eq. 33) transition lengths are great at small |p1| (i.e. large u) and are ∝ 1/|p1| ∝ u. But even in this case Δnunn|   un.
  40. Belinsky, Khalatnikov, Lifshitz (1970), Appendix A
  41. a b c d Belinsky, V.A. (1970). Zhurnal' Eksperimental'noy i Teoreticheskoy Fiziki. 59: 314. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  42. Note that this metric allows arbitrary transformations of type ξ′ + z″ = f1 (ξ + z), ξ′ - z′ = f2 (ξ - z), xa = fa (x1, x2).
  43. The equation   is a direct result of (eq. 97-99) if   or  . The case   does not require a special treatment: it can be shown that the spacetime metric in this case converges (in first approximation) to Galilean.
  44. Einstein, Albert (1937). Journal of the Franklin Institute. 223: 43. {{cite journal}}: Missing or empty |title= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)
  45. Landau, Lifshitz (1988), Classical Field Theory, Ch. 101
  46. It is possible to look for a solution in the form of Fourier integrals; this issue has not been studied in detail. Therefore, BKL do not require Fourier series decomposition as a mandatory condition for the coordinate dependence of functions σ and φ
  47. Squared H terms in (eq. 103) result only in small (~ 1/ξ) corrections in σ and φ. Calculation with cubic terms leads to appearance of a weak dependence of A, B from ξ that can be presented as an appearance of logarithmic phases in the oscillating factors in (eq. 115). These calculations for the case ρ = 0 are given in Belinsky, Khalatnikov (1970), Appendix B (cf. the analogous situation for homogeneous models, Belinsky, Khalatnikov, Lifshitz (1970), Appendix B).
  48. The regular decomposition of the general solution of Einstein equations contains (in addition to the four three-dimensional functions) three independent functions of two coordinates (cf. Petrov, Alexey Z. (1969), Einstein spaces, Oxford: Pergamon Press {{citation}}: Cite has empty unknown parameter: |1= (help), Chapter 40; Lifshitz, Khalatnikov (1963), Appendix A)
  49. Misner, Ch. W. (1969) Phys. Rev. Lett., 22, 1071
  50. Landau, Lifshitz (1988), Classical Field Theory, Ch. 103-105


  • Khalatnikov, Isaak M.; Kamenshchik, Aleksander Yu. (2008), Lev Landau and the problem of singularities in cosmology, To be published in the journal Physics — Uspekhi, {{citation}}: Italic or bold markup not allowed in: |publisher= (help).
  • Bini, Donato; Cherubini, Christian; Jantzen, Robert (2007), The Lifshitz-Khalatnikov Kasner index parametrization and the Weyl Tensor, to appear in the Proceedings of the First Italian-Pakistan Workshop on Relativistic Astrophysics which will be published as a special issue of Nuovo Cimento B,
  • Heinzle, J. Mark; Uggla, Claes; Röhr, Niklas (2007), The cosmological billiard attractor,