M-field I_p can be expressed in U_N via ^NC_M scalar fields i_ij..mp = e^ij...m¿I_p though the normalisation condition renders one such field redundnat apart from sign. In the mapspace ¦^Ñ-1(I_p) is everywhere the map pseudoscalar.
    In U_N the projector 1-tensor ¯ can be expressed with regard to a universal basis as N² scalar fields although with regard to an extended tangential frame at p it can be regarded as an identity matrix having the last N-M terms in the lead diagonal zeroed.

    The Riemann curvature tensor c_paÙb, as a bivector valued function of bivectors, constarined within and upon I_p requires ^MC₂² = ¼M²(M-1)² scalar fields in theory although the Bianchi symmetries reduce the generality considerably. At a fixed given p Î C_M,  mapping bivectors in I_p to bivectors in I_p induces the obvious C_x(aÙb) = ¦^-Ñ_p(c_{p¦^Ñp(aÙb)}) 2-tensor acting in the mapspace which we can represent with ¼M²(M-1)² scalar fields C^kl_ij º e^kl¿C_x(e_ij)

   
    The alternate view of the Riemann curvature as a bivector specific 1-tensor , or a (1;3)-tensor antisymmetric in two of its arguments, provides the alternate coordinate form C^k_ijl º e^k¿(C_x(e_ij)×e_l)


     c_paÙb and C_x(aÙb) are alternate representations of the same "thing", although c_p defined over subspace I_p of U_N can more readily be "extended" out of C_M.     We can contract the curvature tensor to obtain the Ricci 1-tensor)
    c_pb º Ñ_a¿c_paÙb = å_i=1⁴ eⁱ¿c_peiÙb = å_i=1⁴ eⁱ¿ (( Ð_¯(b)w_ei) - ( Ð_¯(ei)w_b) + (w_b×w_ei))    , a point-dependant symmetric 1-vector-valued function of 1-vectors . Note that the divergence is with respect to direction rather than a point and is computable entirely "at" x.
    The mapspace representation of the Ricci 1-tensor consists of M ² scalars C^k_j º e^k¿ C_x(e_i)   = å_j=1^M C^kj_ij or with C_ij º å_k=1^M C^k_ijk which itself contracts to the scalar curvature R = å_i=1⁴ Cⁱ_i .
    The mapspace Einstein tensor is given by M² scalar fields   G^k_j º C^k_j - ½R .

Streamline Coordinates


Introduction
The following is a more traditional approach to manifold calculus which we include for completeness. This section can safely be skipped by readers interested only in multivector methods.

    Suppose we have a 1-field v_p defined over M-curve C_M such that v_p Î I_p . It follows from a mathematical result concerning the unique solvability of ordinary differential equations that, provided v_p is sufficiently smooth and is nowhere 0-valued, for every p₀ÎC_M there is a unique "tangent matcher" path in U^N for v_p :
    C_{i p} = { p(t) : t Î [t₀,t₁] }  [ where t₀<0<t₁ ] with p(0)=p₀ ; p(t) Î C_M and  (¶/¶t)p(t) = v_p(t) " t Î [t₀,t₁] . ( p(0)=p₀ ensures that the parameterisation is unique up to choice of domain range [t₀,t₁])
    Such paths are called integral curves or streamlines for v_p. We will use both terms interchangeably here. Streamlines cannot intercept eachother (other than at points with v_p=0 which we have disallowed) . Since every pÎC_M has such a curve, the set of all integral curves for each v_p (known as a congruence for v_p) "fills" and effectively defines C_M.
    It is important to note that the 1-curve streamlines of field v_p are independant of the magnitude of v_p, (given v_p² ¹ 0) since it is the unit tangent that defines a 1-curve.

    Suppose now we have M not-necessarily mutually orthogonal unit 1-fields u_{i p} defined over M-curve C_M satisfying
    u₁(p)Ùu₂(p)Ù..u_M(p) = a_pI_p     " p Î C_M where scalar field a_p > 0 " p .
    An obvious example is u_{i p}=¦^Ñ_x(e_i)= h_ip  where C_M=¦(M_ap) is a point embedding specification for C_M.

    Given a general multivector-valued function F(p) defined over C_M , u_{i p} generates at each pÎC_M a specific "streamlined" derivative of F which is the tangential derivative within the 1-curve u_{i p} streamline. We will write this as
    d_li(F(p)) º  ¶F(p_i(l_i))/¶l_i . We will sometime ommit the brackets and write d_liF(p).
    If we so differentiate the identity map F(p)=1(p)=p we recover u_{i p} and, indeed, one can think of 1-field u_{i p} as being the streamlined derivative operator field d_lip . We will accordingly use   the symbol Ð_{¯(ui p)} interchangeably with d_li .

    We introduce the notation  _Dli¬ⁱ p to indicate the point p_i(Dl_i) Î C_M reached by moving from p by arclength Dl_i along the l_i (ie. the u_{i p}) streamline through p.
     _Dli¬ⁱ p º ò_{Ci p [0,Dli]} 1 dp = ò₀^Dl_i p_i'(l_i) dl_i
    We can extend  _Dli¬ⁱ  to act on general multivector fields by  _Dli¬ⁱ F(p) º F( _Dli¬ⁱ p) .
    This defintion is integrative, good for all Dl_i within the neighbourhood of ¦ applicability. For small Dl_i=e we can apply Taylor's Theorem to the streamline l_i-parameterisation and obtain F( _e¬ⁱ p) = e^ed_liF(p) and in particular [  Set F(p)=p ] we have  _e¬ⁱ p = e^ed_li(p) which we can write as
     _e¬ⁱ  = e^ed_li .

    We might be tempted to try to parameterise C_M using y:M_ap®C_M defined by
    y(Dl₁, Dl₂, ..Dl_m, q₀) º  _Dlm¬^M  _Dlm-1¬^M-1 ... _Dl2¬²  _Dl1¬¹ q₀ =_( )=  _Dlm¬^M ( _Dlm-1¬^M-1 (... _Dl2¬² ( _Dl1¬¹ q₀))..) .
    [   That is, starting from a fixed reference point q₀ÎC_M we follow the v₁(p) stream line for arclength Dl₁ to a new point  _Dl1¬¹ q₀ from which we follow the v₂(p) streamline for arc length Dl₂ to  _Dl2¬²  _Dl1¬¹ p₀ and so forth. After M such steps we arrive a final point q_MÎC_M which we can associate with parameters (Dl₁,..,Dl_m). ]
    Any M scalars (sufficiently small that we stay within the neighbourhood) interpreted in this way define a point in C_M, but not necessarily uniquely. If it is possible to reach the same point by more than one such route, then the mapping is noninvertible and does not provide a true coordinate system for C_M.
    Only if the d_li (and hence the  _Dli¬ⁱ ) everywhere commute with eachother do the l_i provide a true coordinate system.
[ Proof : ...  .]

    ( _e¬ⁱ  ×  _e¬^j ) is a 1-field defined over C_M returning a U_N 1-vector ( _e¬ⁱ  ×  _e¬^j )(p) = (e^ed_li × e^ed_lj)(p) = ... = e²(d_li×d_lj)(p)   + _O(e³) .
    Taking e®0 we can visualise  (d_li×d_lj)(p) º [ u_{i p},u_jp ](p) as the 1-vector difference at p between following i and then j as opposed to j and then i streamlines by small arclength e.

Lie Derivative
    We make very little use of Lie derivatives in this work. They are described here only for completeness. This section can safely be skipped by readers interested only in multivector methods.

Lie Bracket

    The Lie Bracket is a highly general mathematical construct. With regard to 1-fields in an M-curve we define it by
    [u_{i p},u_jp](v) º  d_li(d_lj[v]) -  d_lj(d_li[v]) = 2(d_li×d_lj)v
which we can write as [u_{i p},u_jp] = d_liu_jp - d_lju_{i p} .
    The use of × above is semantically legitimate if we extend the definition of commutator multivector product × to operators in the obvious appropriate way.

Lie Drag

     Any one of our u_{i p} generates a congruence for C_M so let us fix on u_{i p} as a "preferred" congruence with associated dervivative field d_li . Consider two points p₀ and p₀ in C_M with p₀= _Dli¬ⁱ p₀ for small Dl_i>0. Clearly p₀= _-Dli¬ⁱ p₀ .
    For j¹i, the specific u_jp streamline through p₀ (i¹1) at p with associated directed derivative d_lj generates a distinct path p_j^*(l_j) =  _-Dli¬ⁱ p_j(l_j) passing through point p₀ =  _-Dli¬ⁱ p₀ known as the Lie drag of u_jp . [  Here ^* indicates "new" or "modifed" rather than a multivector dual ] . This path need not be the streamline at p₀ of any of our u_{k p} 1-fields and, in particular, it need not be the l_j streamline there .
    It is  however, a streamline having at p₀ a tangent vector written (ⁱÝ_-Dli(u_jp))(p₀) fully within I_p0 and an associated streamlined differential operator d_lj^* = (ⁱÝ_-Dli(d_lj))(p₀)  .
    Though d_lj^* is not one of the d_lk, it is not general. Because of its construction it commutes with d_li (ie. d_lj^*(d_li(F))) = d_li(d_lj^*(F))) . [ Proof : ....  .]
    We refer to     (ⁱÝ_-Dli(u_jp))(p₀)     ( or (ⁱÝ_-Dli(d_lj))(p₀) ) as the Lie drag of u_jp ( or d_lj ) from p₀ to p₀=  _-Dli¬ⁱ p₀ .
    We define the Lie drag of a scalar field f(p) defined over C_M by (ⁱÝ_-Dli(f))(p) º f( _-Dli¬ⁱ p) so that (ⁱÝ_e(f))(p) = e^ed_lif .                                              7

Lie Derivative
    The 1-vector Lie derivative of a scalar field is simply £_dlif(p) = d_lif(p) .

    Since the Lie drag of u_jp (or d_lj) from p₀ to p₀ can be meaningfully subtracted from u_{i p} (or d_li) at p₀ . The Lie dervivative at p₀ of u_jp with regard to 1-field u_{i p} is an operator defined (by its action on a general 1-field F) as
    ( £_dli(d_lj))F(p) º Lim_{Dli ® 0} [ ( (ⁱÝ_-Dli(d_lj))(p₀)(F(p)) - d_lj(F(p)) )  / Dl_i ] ï_p0 = Lim_{Dli ® 0} [2(d_li×d_lj)F  ï_p0 ]
[ Proof : d_lj^*F ï_p0 = d_lj^*F ï_p0 - Dl_i(d_li(d_lj^*(F))))ï_p0 + _O(Dl_i²) = d_lj^*F ï_p0 + Dl_i(d_li(d_lj(F))ï_p0 + _O(Dl_i²) - Dl_i(d_lj^*(d_li(F))ï_p0 + _O(Dl_i²)
    Hence Lim_{Dli ® 0} [ ( (ⁱÝ_-Dli(d_lj))(p₀) - d_lj )F  / Dl_i ] ï_p0 = Lim_{Dli ® 0} [((d_lj^* - d_lj )F  ï_p0 ) / Dl_i ] = Lim_{Dli ® 0} [(d_li(d_lj(F)) - d_lj^*(d_li(F)) ) ï_p0 ] which provides the result assuming d_lj^* can be safely replaced in the limit by d_lj  .]
    Omitting the brackets , we thus have the 1-vector operator identity
    £_{ui p}u_jp = £_dlid_lj = 2 d_li×d_lj = [ u_{i p},u_jp ] .

Covariant Frame
    Let us now identify the spaces U_N and V_N and adopt fixed frames {e_i} and {eⁱ} in both. With regard to ¿ within V_N , {eⁱ} is not a reciprocal frame for {e_i}. Rather, we need the point-dependant (nonorthormal) covariant frame {e ⁱ º g_x^-1(eⁱ)} which itself has ¿ reciprocal frame {e_i º g_x(e_i)}
[ Proof :   e ⁱ¿e_j = g_x(g_x^-1(eⁱ))e_j = eⁱ¿e_j     Also e ⁱ¿e_j = g_x^-1(eⁱ)¿g_x(e_j) = g_x(g_x^-1(eⁱ))¿e_j = eⁱ¿e_j  .]
    For brevity we ommit an x subscript from e_i , e ⁱ , and ¿ . The underline serves to remind us of the point-dependance.

Covectors

    Any "inner product"  ¿ is a (0;2)-tensor (a potentially point-dependant bilinear scalar-valued function of two 1-vectors in V^M) . For any 1-field ( (1;0)-tensor) v_p , ¿ induces a scalar-valued directional function ( a (0;1)-tensor)     v_p(a) º a¿v_p     known as a covector field. We can recover v_p at a given p from its covector as v_p = Ñ_av_p(a) . A vector and its covector are essentially alternate representations of the same "thing". If we "feed" v_p "to" its covector we get "squared magnitude" 0-tensor v_p(v_p) = v_p¿v_p
    For ¿ in Â_N , if 1-vector v_p is regarded as a 1×N matrix "column vector" , then its covector is the N×1 matrix "row vector" (or transpose) v_p^T . For ¿ in Â_p,q , we take the transpose of the conjugate v_p^† .
    With regard to ¿, we will here define the 1-covector of v_p = å_i=1^N  vⁱe_i as the 1-vector v_p^£ º  å_i=1^N vⁱe_i so that v_p^£ ¿₊ a =    v_p¿a .
    In a Euclidean space, v_p^£ = v_p .
    (1;0)-tensor v_p has "contravariant" coordinate representation å_ivⁱe_i where vⁱ º eⁱ¿v_p .
    (0;1)-tensor v_p has "covariant" coordinate representation å_iv_ie ⁱ where v_i º v_p(e_i) = e_i¿v_p and e ⁱ is the 1-covector of e_i which we associate with the 1-vector e ⁱ.

Parallel transport

    Suppose we were to drag heavy iron girders around the (assumed perfectly spherical) Earth for a bet. We might start at the North Pole, with two girders girder (assumed straight and with ends clearly distinguished) lying "horizontal" on the ground ; perpendicularto a "vertical" flagpole we assume to mark the Pole and parallel to eachother.
    Since the globe is spherical, the girder is actually balanced tangentially on a single point but the radius is so vast that the ground is for all girder-related purposes locally flat. Both gripping one end we start to drag the girder along a meridian, always pulling due South, that is, allowing no rotation in the local "horizontal" tangent plane.  When we reach the equator, or girder has retained its length and is still facing South (local coordinates) but, speaking 3-dimensionally, it is now parallel to the flagpole we leftbehind. If we then push the second girder sideways from the North Pole without rotation the girder maintains an East-West alignment. Once we reach the equator (at a different point to previously) we then drag the girder "lengthwards" ¼ of the way round the equator to join the first girder, to which it is now perpendicular.

    Moving a 1-vector along a curve while keeping it "held within" an M-curve in this way is known as parallel transport in the M-curve. The example shows that it preserves length but not direction. If we parallel transport a 1-vector from p to q within an M-curve then the result is dependant on the path taken.
    Parallel transport is invoked to provide a notion of "parallelism" for 1-vectors in different tangent spaces (ie. 1-vectors "at" different points of an M-curve) that does not require a higher-dimensional embedding space. The "distance" between two points on a manifold can be meaningfully defined as the arclength of a minimial-arclength path joining the points (a geodesic), but to compare distant directions we need for every qÎC_M a function ­_q^p C(a) Î I_p giving the tangent vector at pÎC_M "parallel" to aÎI_q at q ÎC_M ; as obtained by "parallel transportation" of a "along" a given path C from q to p (eg. along a minimal length geodesic).

    Let i­Dlip0(a) denote the parallel transport of a from p0 for arclength Dli along the ui p streamline.
    We obviously want i­Dlip0(a)ÙI Dli¬i p0 = 0 and we might therefore think of defining i­Dlip(a) = ¯I Dli¬i p0(ujp)~ whenever ¯I Dli¬i p0(ujp) is nonnull but instead we simply postulate a 1-vector i­Dlip(a) within I Dli¬i p0 that satifies i­Dlip(a)2 = a2 .   We will henceforth assume a to be a unit vector.

    The normalisation condition for i­Dlip(a) means that the right-connection ba,p(Dli) º a-1i­Dlip(a) for i­Dli at p satisfies ba,pba,p§ = 1 and is thus a 1-rotor expressable as a 2-spinor ewui p a p(Dli) , where wui p a p(Dli) = cos-1(a-1¿i­Dlip(a) (a-1Ùi­Dlip(a)~ is a UN 2-blade. [  we assume Dli is sufficiently small and CM sufficiently smooth that the subtended angle is comfortably below ½p ]  

    We would like the angle subtended by two 1-vectors parallelly transported along the same streamline to remain the same, ie. ⁱ­^e_p(a) ¿ ⁱ­^e_p(b) = a¿b and we can acheive this by insisting that u_{i p}¿w_{ui p a p}(e) = _O(e²) .
    ⁱ­^e_p(a) º ae^{w_{ui p a p}(e)} = a + a¿w_{ui p a p}(e) + _O(e²)
where for given scalar e, w_{ui p a p}(e) is a pure U_N-bivector 2-field defined at every pÎC_M for all aÎI_p .
    Parallel displacement along the i^th streamline thus preserves orthonormality and corresponds to a Lorentz rotation having U_N rotor R_{p i}(e) which we assume to tend to a well defined limit as e®0. M such rotor fields fully define parallel transport and effectively define C_M.

Linear Connection

    g_p(u_jp,d_li,e) º u_jp - ⁱ­^e_p(u_jp) = -u_jp¿w_{ui p ujp p}(e)  + _O(e²)
      is a 1-vector lying in general neither in I_p nor I _e¬ⁱ p. It cannot be properly regarded as a tangent 1-field of the M-curve (traditionally speaking, it is not a "tensor").
    We make the continuity assumption that Lim_{e ® 0}[ _e¬ⁱ u_jp - ⁱ­^-e _e¬ⁱ p( _e¬ⁱ u_jp)] = -g_p(u_jp,u_{i p}). [  We use the symbol g here in accordance with traditional GR use of G for the gravitational connection acting in the mapspace. Conflict with the widespread use in multivector literature of g for a (typically Â_1,3) fiducial frame is unfortunate. ]
    Taking g_p(u_jp,u_{i p}) º Lim_{e ® 0}[e^-1g_{j p}(u_{i p},e) ] we obtain the connection, the 1-vector component of d_li u_jp ï_p that exists outside I_p.
    g_p(u_jp,u_{i p}) = -u_jp¿w_{ui p ujp} = w_{ui p ujp} ×u_jp .
    g_p(u_jp,u_{i p}) is orthogonal to both u_{i p} and u_jp at p .

    A linear connection (aka. affine connection) derives from an assumed bilinearity of g_p which can then can be fully defined by M² 1-fields   g_p( h_jp , h_ip ) which then effectively define ­_q^p C(a). Indeed, some authors define ⁱ­^Dl_i as the consequence of a given linear connection.
    We say the connection is symmetric if g_p(u_{i p},u_jp) = g_p(u_jp,u_{i p}) .

    Suppose now that the u_{i p} are everywhere mutually orthogonal so that {u_{i p}} provides an orthonormal frame for I_p with inverse frame {uⁱ_p}. Because w_{ui p ujp} is a 2-blade perpendicular to I_p, we have w_{ui p ujp} ×u_{i p} = 0     i¹j . Thus we can define M pure bivectors w_{ui p p} º å_j=1^M w_{ui p ujp} which satisfy g_p(u_jp,u_{i p}) =  w_{ui p p}×u_jp
    The connection is symmetric if u_jp¿w_{ui p p} = u_{i p}¿w_{ujp p}     " 1£i,j£M .
    A linear connection arises if w_{ui p a} is linear in a.
[ Proof : Since the u_{i p} are orthogonal g_p((å_j=1^M b^ju_jp),(å_i=1^M aⁱu_{i p})) = (å_j=1^M b^ju_jp)¿w_{ui p åi 1M=} aⁱu_{i p})}
= (å_j=1^M b^ju_jp)¿(å_i=1^M aⁱw_{ui p ui p} ) = å_i=1^M å_j=1^M aⁱb^j(u_jp¿w_{ui p ui p} ) = å_i=1^M å_j=1^M aⁱb^jg_p(u_jp,u_{i p})  .]

Within the mapspace
    Parallel transport in the mapspace ⁱ­^e_x(b) º ¦^Ñ_x+eei^-1( ⁱ­^e_¦(x)(¦^Ñ_p(b)) ) preserves ¿ instead of ¿.
[ Proof : ⁱ­^e_p(a) ¿_x+eei ⁱ­^e_p(b) = ⁱ­^e_p(¦^Ñ_x(a))¿ⁱ­^e_p(¦^Ñ_x(b)) = ¦^Ñ_p(a)¿¦^Ñ_x(b) = a¿_xb  .]

    Let b be a map direction at xÎM_ap with corresponding C_M direction a=¦^Ñ_x(b) at p=¦(x). It may not make any geometrical "sense" to subtract a map 1-vector "at" x+ee_i from one "at" x but we can still do it!
    G_x(b,e_i,e) º b - ⁱ­^e_x(b) = b - ¦^Ñ_x+eei^-1(ⁱ­^e_p(¦^Ñ_x(b))) = b - ¦^Ñ_x+eei^-1(¦^Ñ_p(b) - g_p(¦^Ñ_p(b),d_li,e))
is, for given e_i and scalar e,  a map-1-vector-valued function of map-1-vector b, linear in b if the C_M connection g_p is linear. We then have G_x(e_i,e_j,e) = å_k=1^M  G^k_ijx(e)e_k . Letting   G_x(b,e_i) =   Lim_{e ® 0}(e^-1G_x(b,e_i,e)) we achieve an affine connection for the mapspace, symmetric if g is.
    G_x(a,b) = ¦^-Ñ_x(¦^Ñ2_x(a,b))
[ Proof :   See General Relativity Chapter  .]
    We can think of G as a rule for moving map 1-vectors around. In the chaper on General Relativity we derive an expression for G in terms of the metric.

Directed Coderivative

    The directed coderivative of b(p) with respect to streamline ui p at p is defined by parallel transporting b( e¬i p) back from  e¬i p to p and there subtracting b and dividing by e, in the limit as e®0. It is thus a 1-vector within Ip.
    Ðui p(b) = dlib + gp(b,dli) = dlib + wui p ujp p×b
[ Proof : Ðui p(b(p)) º Lime ® 0[ i­-e e¬i p(b( e¬i p)) - vj(p) / e ]
    = Lime ® 0[ ( i­-e e¬i p(b( e¬i p)) - b( e¬i p)) + ( b( e¬i p) - b(p)) / e ] = gp(b,dli) +dliujp  .]
    In the limit, since li is  the natural (proper) parameterisation, we can consider dli to be equivalent to the N-D directional derivative Ðui p
    Accordingly we can define a directed coderivative operator
    Ða º Ða + wa×
for a given direction aÎIp which when applied at p to a 1-field over and within CM returns a 1-vector in Ip.
    We then have
    i­Dlip(ujp) = eDliÐli e¬i ujp = eDliÐli eDlidli ujp .
[ Proof : i­Dlip(ujp) = vj( Dli¬i p) + DliÐlivj( Dli¬i p) + Dli2Ð (ui p)2vj( Dli¬i p) + ...  .]

    Ð_a can be regarded as a differential over and "within" C_M, although its component parts Ð_a and w_a× cannot.

    Ð_a generates a 1-vector codel-operator
    Ñ_p º å_k=1^M u^k_pÐ_{uk p} = Ñ_[CM] + å_k=1^M u^k_p(w_{uk p}×) .

Within the mapspace
    Ð_a induces a differential operator within M-D parameter space M_apÌV_M
    Ð_a^ð(B(x)) = ¦^-Ñ_p(Ð_¦^Ñp(a)(b(p)))     where b(p) º B(¦^-1(p)) .
    Ð_a^ð(v_p) = Ð_a(v_p) + G(a,v_p)     where G is linear as a consequence of the linearity of ¦^-Ñ_p .
    In particular,
    G_p(e_j,e_i) º Ð_ei^ðe_j = å_k=1^M G^k_jie_k     where point-dependant scalar G^k_ji is known as a Christoffel symbol.

Geodesics
    We say a 1-field ui p is geodesic if i­Dlip(ui p) = ui p( e¬i p)     " pÎCM, Dli
where Dli is assumed small enough to remain in the neighbourhood. Geodesic fields are thus invarient under parallel transport and satisfy Ðui pui p = 0.
    The physical interpretation of geodesics are the potential trajectories of "free falling" particles and (assuming a natural streamline parameterisation pi'(li)2 = 1) we can then regard the
    pi"(li) + g(pi'(li),pi'(li)) = 0
derived from Ðui p(ui p) = 0 as relating the "acceleration" of a "particle" to its velocity and the local "geometry" g.

    An alternate definition of a geodesic is the "shortest subcurve between two points". If we fix two points p and q on a geodesic streamline then the arclength from p to q is stationary (minimal) for small changes in the path from p to q that keep it within C_M. The equivalence of the two defintions is intuitively reasonable and the proof (ommitted here) arises from fairly straightforward calculus

Curvature
[Under Construction]

1-Curvature
    The traditional curvature of a 2D 1-curve p(t) (ie. a 1-curve confined to a 2-plane ) is defined as
    C(t) º Lim_{d ® 0}[ d^-1 q_Ð(p'(t+d) , p'(t)) ] where t is the natural parameterisation - ie. the instantaneous angular tangent change per unit arclenth. C(t)^-1 is known as the raduius of curvature and coresponds to the radius of the circle matching the 1-curve to second order at p(t) . For an N-D 1-curve we have 1-curvature p"(t) which is a 1-vector orthogonal to p'(t) ("normal" to the 1-curve) as a consequence of the constancy of p'(t)² . Its magnitude C(t) is the curvature of the 1-curve confined in the limit to 2-plane p(t).
    For a 3D 2-curve we have tangent unit vectors h₁,h₂ and normal n=h₁×h₂  forming the SYMBOL">t) is the curvature of the 1-curve confined in trihedron) at p. Taking p as the origin, the induced coordinates (x¹,x²,x³) satisfy
    x³ = ½ (¶²x³/¶(x¹)²) x¹²  + (¶²x³/¶x¹¶x²) x¹x² + ½ (¶²x³/¶x²¶) x² |²
    By rotation about e₃ we can diagnonalise this form as
    x³ = ½ (k₁(x¹)² + k₂(x²)²)     where k₁,k₂ are known as principle curvatures.
    If the surface  is parameterised as p(l₁,l₂) with h_1p  = ¶p/¶l₁ , h_2p  = ¶p/¶l₂ then
    ¶²p/¶l_i¶l_j º ¶ h_ip /¶l_j = å_k=1² G^k_ij h_kp  + b_ij h_3p      i,j,k Î {1,2}
    ¶ h_3p /¶l_i = - å_{k l=1}² g^klb_li h_kp      i Î {1,2}

2-Curvature
    The second directional coderivative operator Ð_liÐ_lj is not symmetric in i,j.
    We can take the skewsymmetrol of it to obtain the Rieman Curvature operator
    c_{pui pÙujp}(u)   =   c_{pui pÙujp}×u   =   2(Ð_li×Ð_lj)u = ( Ð_¯(li)w_ujp - Ð_¯(lj)w_{ui p} + w_{ui p}×w_ujp) . u
[ Proof :

ÐbÐau	=	( Я(b) + wb×)( Я(a) + wa×)u
	=	( Я(b) + wb×)( Я(a)u + ½(wau-uwa)) = ( Я(b) Я(a)u + ½( Я(b)wau- Я(b)uwa) +wb×( Я(a)u +   ½(wau-uwa))
	=	Я(b) Я(a)u + ½( Я(b)wau - Я(b)uwa + wb Я(a)u - Я(a)uwb ) + ½wb×(wau-uwa)
	=	Я(b) Я(a)u + ½( Я(b)(wau) + wb Я(a)u - Я(a)(uwb) - Я(b)(uwa) ) + ¼( wbwau - wbuwa - wauwb + uwawb )
Þ 2(¶b׶a)u	=	2( Я(b)× Ð¯(a) + Я(b)×(wa×) + (wb×)× Ð¯(a) + (wb×)×(wa×))u
	=	Я(b)(wau)- Я(a)(wbu) + wb Я(a)u -wa Я(b)u + (wb×wa)×u = ( Я(b)wa)u - ( Я(a)wb)u + (wb×wa)×u
	=	( Я(b)wa)×u - ( Я(a)wb)×u + (wb×wa)×u

 .]
    Given a linear connection, the second-derivative curvature operator  c_paÙb(v) is thus a geometric commutatator product with a particular bivector c_paÙb .

    The more general definition of the curvature operator    is 2(Ð_a×Ð_b) - Ð_[a,b] but we are most interested in the case when [a,b]=0 ( ie. Ð_a×Ð_b =0 ) .

    c_paÙb º Ð_¯(b)w_a - Ð_¯(a)w_b + w_b×w_a =_( )= ( Ð_¯(b)w_a) - ( Ð_¯(a)w_b) + (w_b×w_a)     is known as the curvature 2-tensor (aka. Riemann-Christoffel tensor). It is a C_M-point-dependant I_p-bivector-valued function of I_p-bivectors acting as a "second order correction" for the linear approximator ¦(p₀) + ¦^Ñ_p(dp) for ¦ near p₀.

Geometric interpretation
    If [u_{i p},u_jp]=0 so that we have a loop  _-Dli¬^j  _-Dli¬ⁱ  _Dlj¬^j  _Dli¬ⁱ  p = p then  
    c_{pui pÙujp}(u_{k p}) = (Dl_iDl_j)^{-1  j}­^-Dl_j(  ⁱ­^-Dl_i(  ^j­^Dl_j(  ⁱ­^Dl_i(u_{k p}) ))))
which is clearly dependant solely on value of u_{k p} and the geometry of C_M . In particular, it is independant of any differential of u_{k p}.
    We thus have the geometric interpretation of c_{pui p}u_jp)(u_{k p}) as the 1-vector change in u_{k p} resulting from parallel transport of u_{k p} around (the projection into C_M of) a tiny planar loop through p in u_{i p}Ùu_jp ,  divided by the "area" of that loop. This must lie in I_p so we have
    c_{pui pÙujp}(u_{k p}) = å_l=1⁴ c^l_ijku_lp     where scalar c^l_ijk º u^l_p¿c_{pui pÙujp}(u_{k p}) .
[ Proof : ^j­^{-Dl_j i}­^{-Dl_i j}­^{Dl_j i}­^Dl_i(u_{k p}) = e^-Dl_jÐ_lj e^-Dl_iÐ_li e^Dl_jÐ_lj e^Dl_iÐ_liu_{k p} = ...[tedious expansion]... = u_{k p} + 2Dl_iDl_j(Ð_li×Ð_lj)u_{k p} + _O(Dl³)  .]

Symmetries and Bianci Relations
    It is possible to express C_xeij in terms of the g_ij and their derivatives and doing so reveals some more symmetries:
    C^l_kij º e^l¿(C_xeij(e_k)) = eⁱ¿(C_xelk(e_j)) = -e^k¿(C_xeij(e_l))   .
    The last of these reflects the reversability of parallel transport.
    We also have (from the G expression for Cⁱ_jkl) the first Bianci identity
   e_k¿C_xeij + e_j¿C_xeki + e_i¿C_xejk = 0 .
    which can also be expressed as     C_xeij ¿ e_kl = C_xekl ¿ e_ij .
    It follows that a¿C_x(bÙc)) + b¿C_x(cÙa) + c¿C_x(aÙb) = 0     " 1-vector a,b,c Î I_p.
    As a result, the M ⁴ element matrix representation of C_x has only M ²(M ²-1)/12 independant elements (20 out of 256 for N=4).

    Ð_ei^ðC_xejk + Ð_ej^ðC_xeki + Ð_ek^ðC_xeij = 0₂         is known as the second Bianci identity.

    The codivergence of the Ricci tensor Ñ^ð_x¿ c_pa º å_ieⁱ¿Ð_ei c_pa = ½Ð_aC_x
[ Proof : Ñ^ð_x¿ c_pb = Ñ^ð_x¿(Ñ_a¿c_pab)) Þ ... ???  .]

Tortion
    The differential operator Ð_a Ð_¯(b) - Ð_b Ð_¯(a) - 2( Ð_¯(a)× Ð_¯(b)) is known as the tortion operator.   It is 0 for a symmetric connection when we thus have
    £_{ui p}u_jp = Ð_{ui p} Ð_¯(ujp) - Ð_ujp Ð_{¯(ui p)} .
    Suppose now that u_{i p} is a geodesic field and 1-vector b_p is Lie dragged along geodesic u_{i p} congruence so that £_{ui p}b_p = 0 . We must then have
    Ð_{ui p}² Ð_¯(bp) = 2(Ð_{ui p}×Ð_bp) Ð_{¯(ui p)} .
[ Proof :  Ð_{ui p} Ð_¯(bp) = Ð_bp Ð_{¯(ui p)} Þ Ð_{ui p}² Ð_¯(bp) = Ð_{ui p}Ð_bp Ð_{¯(ui p)} = 2(Ð_{ui p}×Ð_bp) Ð_{¯(ui p)} + Ð_bpÐ_{ui p} Ð_{¯(ui p)}   = 2(Ð_{ui p}×Ð_bp) Ð_{¯(ui p)} since u_{i p} geodesic Þ Ð_{ui p} Ð_{¯(ui p)} = 0  .]
    Thus we have a second geometric interpretation of the curvature tensor as the second derivative of a vector dragged along a geodesic confluence, informally: a measure of geodesic "splay".

[Under Construction]

Extremal M-Curves
    A standard problem in classical mechanics is to determine the form of the curve assumed by a rope of length a hanging motionless between two fixed points a distance less than a apart under a uniform gravitational field, which is an extremal 1-curve in N=2 dimensions.
    If we assume the rope to hold a curve shape p(s)=(x(s),y(x)) (the Y axis being gravitationally vertical) that does not loop "over" itself vertically, we have ds² = dx² + dy² so that the length constraint is a = ò₀¹ ds = ò_x0^x₁ dx(1+y'²)^½ where y' º dy/dx.
    The mechanical constraint is that total gravitational potential energy ò₀¹ ds rgy(s)   =   rg ò_x0^x₁ dx y(1+y'²)^½ be minimised, where g is uniform vertical gravitational acceleration and r is the mass of unit length of rope.
    More generally, we might seek to minimise ò_x0^x₁ dx¦(x,y,y') subject to a constraint ò_x0^x₁ dxg(x,y,y') = a.
    A standard approach is to form h(x,y,y') º ¦(x,y,y') + lg(x,y,y') where scalar l is known as a Lagrange multiplier. One then embodies freedom to vary the path by means of two z₀ and z₁ such that y(x₀,z₀,z₁)=y(x₀) ; y(x₁,z₀,z₁)=y(x₁) " z₀,z₁ (boundary condition) ; y(x,0,0)=y(x) ; and y(x,z₀,z₁) is twice continuosuly differentiable in all parameters.
    Forming K(z₀,z₁) = ò_+x0^x₁ dx h(x,y(x,z₀,z₁),y'(x,z₀,z₁)) the extremal condition requires ¶K/¶z₀ and ¶K/¶z₁ to vanish at z₀=z₁=0 and this leads (via integration by parts) to the Euler-Lagrange equation ¶h/¶y = d/dx(¶h/¶y') .
    In the case of the hanging rope, h(x,y,y') = (rgy-l)(1+y'²)^½ and setting l = -rgy₀ we have   h(x,y,y') = rg(y-y₀)(1+y'²)^½ = rg(y-y₀)(1+(y-y₀)'²)^½ for some constant y₀ .
    Setting h = y-y₀ the Euler-Lagrange equation is rg(1+h'²)^½ = d/dx   rghh'(1+h'²)^-½ with first order simplification
    h' rghh'(1+(y-y₀)'²)^-½ - rgh(1+h'²)^½ = b Þ rg(1+h'²)^-½h = b having solution h = b cosh(x-c)/b) so the rope has shape y₀ + b cosh((x-c)/b) where c,y₀ and b are chosen to match the given endpoints.

    We say an M-curve C_M is M-extremal for an action functional ¦_CM:U^N ® U_N if some particular magnitude measure (content, scalar part, square, modulus or whatever) of ò_CM d^M-1p ¦_CM(p) is maximised (or minimised) by C_M in the sense that integrating other any M-curve which deviates only slightly from C_M will produce a result no higher (or lower) than the C_M integral. C_M is a "locally optimal" M-curve for a particular integration ¦_CM. We specify the dependancy of ¦ on C_M to allow action functionals dependant on the geometric properties of the "sampling curve" C_M , such as tangent or normal vectors , to allow "direction dependant sampling" or "velocity dependance".    

Euler-Lagrange Equations
    For M=1 we have a path p(s) 1-extremal for F_C1 = F(s,p(s),p'(s)).

    Hamilton's principle provides that the state of a system characterised at time t by k multivector variables x₁(t),..,x_k(t) varies from time t₀ to t₀ so as to 1-extremise a (usually real scalar) integral measure
    S(t) º S(t₀) + ò_t0^t₀ dt L(t,x₁(t),..,x_k(t),x'(t),...,x_k'(t))     with the action functional L(t,x₁(t),..,x₁'(t)) known as the Lagrangian of the system.
    L is usually assumed to be real-scalar valued and independant of p"(t) and higher derivatives and is traditionally assumed to seperate into "kinetic" and "potential" componenets independent of time and velocity respectively as L(t,p(t),p'(t)) = T(p(t),p'(t)) - V(t,p(t)) but more generally we might postulate a multivector-valued Lagrangian of k multivector-valued variables and their first temporal derivatives L(t,x₁(t),..,x_k(t),x₁'(t),..,x_k'(t)).
    If the Lagrangian L is itself an integral over some spacial M-curve (typically an M=N-1 hypercurve reprenting a contemporal slice of a B_ase space) L = ò_Base(t) d^N-1p L(t,p,p') that spacially integrated function is known as a Lagrangian density. Note that such requires a "velocity" p' be associated with every point p in B_ase(t)   .

Extremal Paths
    A standard approach for M=N=1 to 1-extremise ¦(s,x,x') is to embody freedom to vary the path by means of two scalar parameters z₀ and z₁ such that x(s₀,z₀,z₁)=x(s₀) ; x(s₁,z₀,z₁)=x(s₁) " z₀,z₁ (boundary condition) ; x(s,0,0)=x(s) ; and x(s,z₀,z₁) is twice continuosuly differentiable in all parameters. Forming K(z₀,z₁) = ò_s0^s₁ ds¦(s,x(s,z₀,z₁),x'(s,z₀,z₁)) the extremal condition requires ¶K/¶z₀ and ¶K/¶z₁ to vanish at z₀=z₁=0 and this leads (via integration by parts) to the Euler-Lagrange equation ¶¦/¶x = (d/ds)(¶¦/¶x') .
    In particular cases where ¦(s,x,x') = ¦(x,x') so that ¶¦/¶s=0 and there is no explicit s dependence,   we have d/ds(x'¶¦/¶x' - ¦)= 0 giving a first order differential equation x'¶¦/¶x' - h = constant .
[ Proof :   y" ¶h/¶y' + y'd/dx¶h/¶y' - ¶h/¶x-¶h/¶yy' - ¶h/¶y'y"   =   y'(d/dx¶h/¶y' - ¶h/¶y) - ¶h/¶x   =   0  .]

    A nongeometric generalisation is to extremise ò_s0^s₁ ds¦(s,x₁,x₂,..x_k,x₁',x₂',..x_k') subject to M scalar constraints g_i(s,x₁,x₂,...x_k)=0 i=1,2,..,M (Note the absence of any g dependance on the x_i'). To do so we form
    h(s,x₁,..,_xK,x₁',..,x_k') = ¦(s,x₁,..,_xK,x₁',..,x_k') + å_j=1^M l_j(s)g_j(s,x₁,..,x_k,x₁',..,x_k') where l_j(s) are M arbitary functions of s and obtain Euler-Lagrange formulae ¶h/¶x_i = (d/ds)(¶h/¶x_i')     for i=1,2,..k .

    Generalising geometrically, we have F(s,x₁,x₂,..,x_k,x₁',x₂',..,+x_k') with geometric Euler-Lagrange Equations ¶F/¶x_i =  (d/ds)(¶F/¶x_i')     " i=1,2,..k and first order equation å_i=1^k  (x_i'_*Ñ_xi')F - F = constant (independant of s) if ¶F/¶s = 0. Considering the x_i as seperate grades of a single multivector argument x we can regard the Euler-Lagrange equations as individual coordinate terms of a single geometric equation
    ¶_xF(s,x,x') = (d/ds)(¶_x'F(s,x,x')) where ¶_x = å_ijk.. e^ijk..¶/¶x^ijk.. over all blades comprising x space so that, for example, ¶_x = å_i=1^N eⁱ(¶/¶xⁱ).

Generalised Momentum

    When F is a scalar-valued Lagrangian we have  (x_i'_*¶_xi')L = x_i'_*(¶_xi'L) and multivector m = Ñ_xi'L(x₁,..x_k,x₁',..,x_k) is known as generalised spacial momenta or cononical momenta. Typically x_i' is spacial 1-vector valued and so momenutum is a spacial 1-vector.
    If multivector L is independant of x_i so that ¶L/¶x_i=0 then the i^th Euler-Lagrange equation provides (d/dt) ¶L/¶x_i' = 0 so M_i = ¶L/¶x_i' is constant, ie. unchanging with t,

    Thus the constants of a system are consequences of absent dependendencies (aka. symmetries) in the Lagrangian. Energy is conserved when L depends on t only indirectly via x(t) and it derivatives. Momentum is conserved when L depends on x only indirectly via x'.
    For L(t,x,x') = ½mx'² - f(x,t) we obtain classical momentum Ñ_x'½mx'² = mx' , constant if f(x,t)=f(t).
    Adding an electromagnetic term qc^-1a_x¿x' to L introduces qc^-1a_x to the momentum

    If L(t,x₁(t),..,x_k(t),x₁'(t),..,x_k'(t))   =   L(x₁(t),..,x_k(t),x₁'(t),..,x_k'(t)) so that there is no explicit t dependance then the Hamiltonian H(x₁(t),..,x_k(t),x₁'(t),..,x_k'(t)) º å_i=1^k  (x_i'_*Ñ_xi')L - L     is a constant, the "energy" of the system.
    We can regard the Hamiltonian as generalised temporal momentum. Conservation of energy but varying spacial momentum resulting from only non-relativistic potentials f(x,t)=f(x) .

    But H = - ¶S/¶t .

    Even though, like the Lagrangian, S(t) cannot strictly be regarded as a function of position since it is only defined over our given extremal path, we can nontheless postualte an action field S(t,x) an it can be shown that
    m = (x_i'_*¶_xi')L = Ñ_xiS(t,x_i) and in particular we have the Hamilton-Jacobi equation H(x_i,x_i',t) = -¶S/¶t
[ Proof :     For any ¦(x,x')
    d/dt ( å_j=1^k  x_i' ¶¦/¶x_j' - ¦)   =   å_j=1^k  x_j" ¶¦/¶x_j' + å_i=1^k  x_j' d/dt ¶¦/¶x_j' - ¶¦/¶t - å_j=1^k ¶¦/¶x_jx_j' - å_j=1^k ¶¦/¶x_j'x_j"   =   å_j=1^k  x_j' d/dt ¶¦/¶x_j' - ¶¦/¶t - å_j=1^k ¶¦/¶x_jx_j'   =   å_j=1^k  x_i'( d/dt(¶¦/¶x_j') - ¶¦/¶x_j ) - ¶¦/¶t
      =   å_j=1^k  x_j'( d/dt(¶¦/¶x_j') - ¶¦/¶x_j )     if ¶¦/¶t = 0
      =   0 if d/dt(¶¦/¶x_j') = ¶¦/¶x_j     j=1,2,..,k which are the Euler-Lagrange equations.

    d/dx ( å_i=1^k  y_i' ¶¦/¶y_i' - ¦)   =   å_i=1^k  y_i" ¶¦/¶y_i' + å_i=1^k  y_i' d/dx ¶¦/¶y_i' - ¶¦/¶x - å_i=1^k ¶¦/¶y_iy_i' - å_i=1^k ¶¦/¶y_i'y_i"   =   å_i=1^k  y_i' d/dx ¶¦/¶y_i' - ¶¦/¶x - å_i=1^k ¶¦/¶y_iy_i'   =   å_i=1^k  y_i'( d/dx(¶¦/¶y_i') - ¶¦/¶y_i ) - ¶¦/¶x
      =   å_i=1^k  y_i'( d/dx(¶¦/¶y_i') - ¶¦/¶y_i )     if ¶¦/¶x = 0
      =   0 if d/dx(¶¦/¶y_i') = ¶¦/¶y_i     i=1,2,..,k which are the Euler-Lagrange equations.  .]
    Theoretical physics then becomes a quest for the One True Action Lagrangian density , usually assumed real scalar and generating kinematic equations involving zeroth, first,  and second differentials only. Geometric (multivector-valued) Lagrangians extermised in the sense that each multivector coordinate is stationary under variation. This is an intrinsically nonrelativistic approach in the case of multiple particulate systems since it requires a favoured temporal parametrisation t with L(t,p,p') = L(p,p')  
    As "locally shortest routes", Geodesics are 1-extremals for the scalar path length of ò₀^t ds|d¹p¿g(d¹p)|^½ . Integrating a square root can be messy but fortunately such paths also extremise   ò₀^t ds ds p'(s)¿g_p(s))(p'(s)) where p'(s) denotes the (d/ds)p(s) . _Be Thus we set L(p,p',s) = p'¿g_p(p')|^½ and minimise scalar integral   ò₀^t ds ds L(p(s),p'(s),s) by solving the N Euler-Lagrange equations d/ds(¶L/¶pⁱ') = ¶L/¶pⁱ) for i=1,2,...N .
    L(p,p',s) will generally depend on p via _gp but if _gp has a symmetry such that ¶L/¶pⁱ=0 for a particular coordinate i then ¶L/¶pⁱ' is constant along any geodesic path.
    Geodesic flow can consequently be viewed as arising from a Lagrangian density L(p(t),p'(t)) = |p'(t)¿_gp(p'(t))|^½ . with particles following timelike trajectories that extremise (minimally) their proper time (arclength).
    If t is proper time parameterisation then L(p(t),p'(t)) = |p'(t)²|^½ = 1 along a geodesic.
    Electrodynamical forces due to a 1-vector four-potential a_p (typically with Ñ_p¿a_p=0) are introduced by adding scalar -q p'(t)¿ a_p(t) to the Lagrangian density leading to the Lorentz force law mp"(t) = qf_p.p'(t) where f_p = Ñ_pÙa_p .
[ Proof : L = m(-p'²)^½ - q p'¿a_p Þ Ñ_pL = Ñ_p(m(-p'²)^½ - q p'¿ a_p) = -qÑ_p(p'¿a_p) ; Ñ_p'L = -m(-p'(t)²)^-½ p'(t) - qa_p     so the Euler-Lagrange equation is qp'¿Ñ_pa_p = (d/dt)(m(-p'²)^-½ p' + qa_p ) = mp" + q(p'¿Ñ_p)a_p Þ mp" = q(Ñ_p(a_p¿p') - (p'¿Ñ_p)a_p))   =   q(Ñ_pÙa_p)¿p'  .]

Extremal surfaces
    A more geometric generalistaion is to consider the problem of finding the M-curve of given content maximising a particular boundary or interior integral. We might, for example, seek the loop starting and ending at a given point a and constrained to lie in a given 2-curve (surface) containing a that maximises enclosed content (area) for a given boundary content (pathlength)