diff f6bf760..6fbd3ac tracial_property_gamma_v2.txt
diff --git a/tracial_property_gamma_v2.txt b/tracial_property_gamma_v2.txt
index 869544f..70cb468 100644
--- a/tracial_property_gamma_v2.txt
+++ b/tracial_property_gamma_v2.txt
@@ -54,7 +54,7 @@ Throughout, A is a separable C*-algebra.  A C*-algebra is unital
 unless stated otherwise.  All ideals are closed and two-sided.
 
 --------------------------------------------------
-1.1  Traces and their norm inequalities
+1.1  Traces and their norm ineualities
 --------------------------------------------------
 
 T(A) is the set of tracial states: linear functionals tau: A -> C with
@@ -68,7 +68,7 @@ For tau in T(A) put
     ||a||_{2,tau} := tau(a*a)^{1/2},    ||a||_{1,tau} := tau(|a|).
 
 For a subset X of T(A) write ||a||_{2,X} := sup_{tau in X} ||a||_{2,tau}.
-The following inequalities are used throughout:
+The following ineualities are used throughout:
 
     |tau(b*a)|       <= ||b||_{2,tau} ||a||_{2,tau}     (Cauchy-Schwarz)
     |tau(a b)|       <= ||a|| ||b||_{1,tau}              (duality)
@@ -720,7 +720,7 @@ property,
 
     tau(a q b) <= ||a q||_{2,tau} ||q b||_{2,tau} <= ||a|| ||b|| tau(q).
 
-Applying this map to the inequality (*) and multiplying by the unitary
+Applying this map to the ineuality (*) and multiplying by the unitary
 x*, we obtain
 
     ||Phi(q) - x Phi(q) x*||_{1,tau} <= mu tau(q).  End.  []
@@ -863,9 +863,9 @@ gives, for a unitary x,
       <= 2 sqrt(tau(b)) ||[x, sqrt(b)]||_{2,tau}
       <= 2 sqrt(tau(b) ||b - x b x*||_{1,tau}).                    (#)
 
-Middle inequality: factor |s - t| = |sqrt(s) - sqrt(t)| (sqrt(s) + sqrt(t))
+Middle ineuality: factor |s - t| = |sqrt(s) - sqrt(t)| (sqrt(s) + sqrt(t))
 and use Cauchy-Schwarz; the second squared sum is at most
-2 sum_{s,t} (s + t) ||e_s x e_t||^2 = 4 tau(b).  Last inequality uses
+2 sum_{s,t} (s + t) ||e_s x e_t||^2 = 4 tau(b).  Last ineuality uses
 Lemma 3.2.
 
 For b := b_q, using (mass)-type estimates plus (centrality) of
@@ -1430,7 +1430,7 @@ Proof.  Every tau in T(A) gives a functional [a] |-> d_tau(a) on Cu(A)
 by ERS Prop 4.2 (applied to the lower semicontinuous trace tau (x) Tr).
 Only this direction of the trace-functional correspondence is used.
 
-Let u, v in (A (x) K)_+ satisfy the functional inequalities.  If v = 0,
+Let u, v in (A (x) K)_+ satisfy the functional ineualities.  If v = 0,
 any tau in T(A) gives d_tau(u) = 0, and the faithful stabilized trace
 then gives u = 0, so u <~ v.
 
@@ -1468,7 +1468,7 @@ Define
     I := { (a_l) in l^inf(N, A) : lim_{l to omega} ||a_l||_{2, T(A)} = 0 }
     D := l^inf(N, A) / I.
 
-The tracial inequalities xy, yx, x* all obey the correct 1-Lipschitz
+The tracial ineualities xy, yx, x* all obey the correct 1-Lipschitz
 bound in ||.||_{2, tau}; so I is a closed two-sided *-ideal and D is a
 unital C*-algebra.  Faithfulness of every tau makes a |-> (a, a, ...)
 injective, so we identify A with its image in D.
@@ -1576,7 +1576,7 @@ For example:
     ||[x, a] - [x, b]||_{2, Lambda} <= 2 ||x|| ||a - b||
     sup_{tau in Lambda} |tau((a - b) x)| <= ||a - b|| ||x||.
 
-So the identities and inequalities extend from the dense subset to
+So the identities and ineualities extend from the dense subset to
 every coefficient by continuity.  In particular a vanishing uniform
 2-norm commutator gives exact commutation in D.
 
@@ -1734,7 +1734,7 @@ A (x) K with ||X* b X - v|| < t and set
     w := x* b x = P_L X* b X P_L in M_L(A),   ||w - v|| < t.
 
 Put alpha := ||w - v||, gamma := t - alpha > 0.  In M_L(A)** put
-E := 1_{(3 t, inf)}(v), F := 1_{(t, inf)}(w).  The order inequalities
+E := 1_{(3 t, inf)}(v), F := 1_{(t, inf)}(w).  The order ineualities
 (w - 2 t)_+ >= w - 2 t 1 and w >= v - alpha 1 give
 
     E (w - 2 t)_+ E >= E (v - (2 t + alpha) 1) E >= gamma E,
@@ -1745,7 +1745,7 @@ Neumann algebra and Q a projection with Q a Q >= gamma Q for gamma > 0.
 Then the polar decomposition of a^{1/2} Q has initial projection Q
 (absorbs through Q a Q >= gamma Q) and final projection at most
 supp(a) (its range is inside the range of a^{1/2}).  Applying that
-observation to the two displayed inequalities, and taking normal
+observation to the two displayed ineualities, and taking normal
 traces, yields
 
     d_tau((v - 3 t)_+) <= d_tau((w - 2 t)_+),
@@ -1776,7 +1776,7 @@ Taking traces and using equality of cut ranks:
     tau(p_0 - p_0^2) <= d_tau((w - t)_+) - d_tau((w - 2 t)_+).
 
 Combine with (rank-xfer) and (uniform-cut) to get the stated
-inequalities.  v = 0 is included (the rank approximation then gives
+ineualities.  v = 0 is included (the rank approximation then gives
 tau(b) < delta for all tau, and x = 0 is a witness).  End.  []
 
 --------------------------------------------------
@@ -1808,7 +1808,7 @@ Let p be the class of (p_l) in D.  Since p_l is a positive contraction,
 
 so the projection defect has vanishing uniform 2-norm along omega;
 p = p* = p^2 in D.  For an arbitrary lambda = lim_omega tau_l on
-represented sequences, the first inequality gives
+represented sequences, the first ineuality gives
 
     lambda(p) = lim_{l to omega} tau_l(b_l) = lambda((h - eta)_+).
 
@@ -1889,8 +1889,8 @@ and tau in Lambda, t := tau(u - v).  Since u - v is a projection:
 
     ||u - v||_{2, tau} = sqrt(t),   ||u x u - v x v||_{2, tau} <= 2 sqrt(t).
 
-The second inequality writes the difference as
-(u - v) x u + v x (u - v) and uses the tracial 2-norm inequalities.
+The second ineuality writes the difference as
+(u - v) x u + v x (u - v) and uses the tracial 2-norm ineualities.
 Telescope powers (all contractive):
 
     ||(u x u)^k - (v x v)^k||_{2, tau} <= 2 k sqrt(t)   (k >= 1).
@@ -2168,7 +2168,7 @@ projection e with orthogonality of phi(e) and phi(1_F - e),
 strongly as t -> 0.  Because projections span F linearly, pi_t has a
 pointwise strong limit pi: F -> D**, CP, pi(1_F) = q, and pi takes
 projections to projections with range in qD**q.  As a UCP map into
-that corner, pi is multiplicative: equality in the Schwarz inequality
+that corner, pi is multiplicative: equality in the Schwarz ineuality
 at a projection e (Schwarz applied to e + z x for all z in C) forces
 pi(e x) = pi(e) pi(x) and pi(x e) = pi(x) pi(e); taking linear
 combinations gives multiplicativity.  Also
@@ -2191,7 +2191,7 @@ projection p in D with
     p <= 1_{[eta, 1]}(a) <= pi(e_{11}),
     tau(p) = tau((a - eta)_+)   for every tau in Lambda.
 
-The last inequality follows from pi(e_{11}) = supp(a).  Define on [0, 1]:
+The last ineuality follows from pi(e_{11}) = supp(a).  Define on [0, 1]:
 
     k(s) := min(s / eta, 1),
     g(s) := k(s) / s  on (0, 1],   g(0) := 1 / eta.
@@ -2242,7 +2242,7 @@ commutation with pi(c) give
     ||rho(c) - pi(c)||_{2, tau}^2 <= tau(q - P) <= 1 - tau(h) + eta.
 
 Here q - h is a positive contraction and q - P a projection, and both
-commute with pi(c).  Squared triangle inequality finishes.  End.  []
+commute with pi(c).  Squared triangle ineuality finishes.  End.  []
 
 --------------------------------------------------
 9.3  An observable with exact coefficient moments
@@ -2292,7 +2292,7 @@ Lemma 9.2:
     sum_i beta_i ||T_i(a) - rho_i psi_i(a)||_{2, tau}^2
       <= 4 sum_i beta_i (1 - tau(h_i)) + 2 eta <= 6 eta.      (six)
 
-Second triangle inequality with (three) and (six):
+Second triangle ineuality with (three) and (six):
 
     sum_i beta_i ||a - rho_i psi_i(a)||_{2, tau}^2
       <= 2 (3 eta) + 2 (6 eta) = 18 eta.                      (eighteen)
@@ -2401,7 +2401,7 @@ finite positive measure with mu_a(U) := < E_z(U) a, a >.  Since g_R is
 
 This proves (comm).
 
-Scalar inequalities.  For real s,
+Scalar ineualities.  For real s,
 
     |s - g_R(s)|    <= |s|^4 / R^3,
     |s^2 - g_R(s)^2| <= |s|^4 / R^2,
@@ -2420,7 +2420,7 @@ which follows from the polar decomposition x = u|x| in D**, with y :=
       <= tau(y^2) = tau(|x|).
 
 So (tail) and (fourth) hold without a commutation assumption on b.
-Combining with Prop 9.3 gives the four displayed inequalities.  End.  []
+Combining with Prop 9.3 gives the four displayed ineualities.  End.  []
 
 --------------------------------------------------
 9.5  Local bounded moments in A
@@ -2582,7 +2582,7 @@ norm <= 1.  Write d := d_{N,j}.  Scalar Taylor with integral remainder:
 
     | e^{ix} - 1 - ix + x^2 / 2 | <= |x|^3 / 6      (x in R).
 
-Functional calculus followed by the tracial inequality
+Functional calculus followed by the tracial ineuality
 | tau(b y) | <= tau(|y|) for ||b|| <= 1 (from 9.4 (comm)'s
 polar-decomposition argument) gives
 
@@ -2591,7 +2591,7 @@ polar-decomposition argument) gives
       <= |t|^3 / (6 N^{3/2}) tau(|d|^3)
       <= 3^{3/4} |t|^3 / (6 N^{3/2})                             (taylor)
 
-The last inequality is Holder's inequality for the probability measure
+The last ineuality is Holder's ineuality for the probability measure
 associated with d at tau, together with tau(|d|^3) <= tau(d^4)^{3/4}
 and (row-moments).  Inserting the two conditional moment bounds from
 (row-moments) into (taylor), for every 1 <= j <= N:
@@ -2880,7 +2880,7 @@ Define
     D := A^omega := l^inf(A) / I_T,
     I_T := { (a_n) in l^inf(A) : lim_{n to omega} ||a_n||_{2, T} = 0 }.
 
-The tracial inequalities make I_T a closed two-sided *-ideal; D is a
+The tracial ineualities make I_T a closed two-sided *-ideal; D is a
 C*-algebra.  We regard A as a subalgebra of D through constant
 sequences; simplicity and faithfulness of members of T make this
 injective.
@@ -3299,7 +3299,7 @@ By Hahn-Banach separation there is a nonzero bounded linear functional
 phi on C(T, R) with phi <= phi_on_O on the hull of {L}.  Since O is
 closed under adding ARBITRARY negative functions, phi is positive:
 if not, adding a strongly negative function would push < phi, f >
-arbitrarily far down avoiding the separation inequality.  So phi
+arbitrarily far down avoiding the separation ineuality.  So phi
 normalizes to a probability measure mu on T.
 
 Barycenter.  Let sigma_mu be the barycenter of mu in compact convex T.
@@ -3603,7 +3603,7 @@ Lemma 11.1 for trace comparison and 2-norm detection):
 
 Commutators with the dense sequence (x_k)_k of zero uniform 2-norm
 give exact commutation, so e in E'.  Norm continuity extends the
-trace inequality from the dense self-adjoint unit ball to the whole
+trace ineuality from the dense self-adjoint unit ball to the whole
 ball, and scaling to every self-adjoint element of E.  End.  []
 
 --------------------------------------------------
@@ -3635,7 +3635,7 @@ Define
 These are positive contractions of norm at most one, commuting with E
 and with one another.  For a self-adjoint contraction x in E, the
 element P_{j-1} x is a permitted test at step j (it lies in the
-enlarged algebra and has norm <= 1).  The trace inequality of Lemma
+enlarged algebra and has norm <= 1).  The trace ineuality of Lemma
 14.1 gives, uniformly in tau in Lambda,
 
     tau(P_j x) = tau(P_{j-1} x) - tau(e_j P_{j-1} x)