diff --git a/src/analysis/calculus/mean_value.lean b/src/analysis/calculus/mean_value.lean
index 11a1763623154..2e6590a598b54 100644
--- a/src/analysis/calculus/mean_value.lean
+++ b/src/analysis/calculus/mean_value.lean
@@ -4,6 +4,7 @@ Released under Apache 2.0 license as described in the file LICENSE.
 Authors: Sébastien Gouëzel, Yury Kudryashov
 -/
 import analysis.calculus.local_extr
+import analysis.convex.slope
 import analysis.convex.topology
 import data.complex.is_R_or_C
 
diff --git a/src/analysis/convex/function.lean b/src/analysis/convex/function.lean
index 339224bff2a8d..d0a10b5dee864 100644
--- a/src/analysis/convex/function.lean
+++ b/src/analysis/convex/function.lean
@@ -485,110 +485,6 @@ lemma concave_on_iff_div {f : E → β} :
   → 0 < a + b → (a/(a+b)) • f x + (b/(a+b)) • f y ≤ f ((a/(a+b)) • x + (b/(a+b)) • y) :=
 @convex_on_iff_div _ _ (order_dual β) _ _ _ _ _ _ _
 
-/-- For a function `f` defined on a convex subset `D` of `𝕜`, if for any three points `x < y < z`
-the slope of the secant line of `f` on `[x, y]` is less than or equal to the slope
-of the secant line of `f` on `[x, z]`, then `f` is convex on `D`. This way of proving convexity
-of a function is used in the proof of convexity of a function with a monotone derivative. -/
-lemma convex_on_of_slope_mono_adjacent {s : set 𝕜} (hs : convex 𝕜 s) {f : 𝕜 → 𝕜}
-  (hf : ∀ {x y z : 𝕜}, x ∈ s → z ∈ s → x < y → y < z →
-    (f y - f x) / (y - x) ≤ (f z - f y) / (z - y)) :
-  convex_on 𝕜 s f :=
-linear_order.convex_on_of_lt hs
-begin
-  assume x z hx hz hxz a b ha hb hab,
-  let y := a * x + b * z,
-  have hxy : x < y,
-  { rw [← one_mul x, ← hab, add_mul],
-    exact add_lt_add_left ((mul_lt_mul_left hb).2 hxz) _ },
-  have hyz : y < z,
-  { rw [← one_mul z, ← hab, add_mul],
-    exact add_lt_add_right ((mul_lt_mul_left ha).2 hxz) _ },
-  have : (f y - f x) * (z - y) ≤ (f z - f y) * (y - x),
-    from (div_le_div_iff (sub_pos.2 hxy) (sub_pos.2 hyz)).1 (hf hx hz hxy hyz),
-  have hxz : 0 < z - x, from sub_pos.2 (hxy.trans hyz),
-  have ha : (z - y) / (z - x) = a,
-  { rw [eq_comm, ← sub_eq_iff_eq_add'] at hab,
-    simp_rw [div_eq_iff hxz.ne', y, ←hab], ring },
-  have hb : (y - x) / (z - x) = b,
-  { rw [eq_comm, ← sub_eq_iff_eq_add] at hab,
-    simp_rw [div_eq_iff hxz.ne', y, ←hab], ring },
-  rwa [sub_mul, sub_mul, sub_le_iff_le_add', ← add_sub_assoc, le_sub_iff_add_le, ← mul_add,
-    sub_add_sub_cancel, ← le_div_iff hxz, add_div, mul_div_assoc, mul_div_assoc, mul_comm (f x),
-    mul_comm (f z), ha, hb] at this,
-end
-
-/-- For a function `f` defined on a subset `D` of `𝕜`, if `f` is convex on `D`, then for any three
-points `x < y < z`, the slope of the secant line of `f` on `[x, y]` is less than or equal to the
-slope of the secant line of `f` on `[x, z]`. -/
-lemma convex_on.slope_mono_adjacent {s : set 𝕜} {f : 𝕜 → 𝕜} (hf : convex_on 𝕜 s f)
-  {x y z : 𝕜} (hx : x ∈ s) (hz : z ∈ s) (hxy : x < y) (hyz : y < z) :
-  (f y - f x) / (y - x) ≤ (f z - f y) / (z - y) :=
-begin
-  have h₁ : 0 < y - x := by linarith,
-  have h₂ : 0 < z - y := by linarith,
-  have h₃ : 0 < z - x := by linarith,
-  suffices : f y / (y - x) + f y / (z - y) ≤ f x / (y - x) + f z / (z - y),
-  { ring_nf at this ⊢, linarith },
-  set a := (z - y) / (z - x),
-  set b := (y - x) / (z - x),
-  have heqz : a • x + b • z = y, by { field_simp, rw div_eq_iff; [ring, linarith] },
-  have key, from
-    hf.2 hx hz
-      (show 0 ≤ a, by apply div_nonneg; linarith)
-      (show 0 ≤ b, by apply div_nonneg; linarith)
-      (show a + b = 1, by { field_simp, rw div_eq_iff; [ring, linarith] }),
-  rw heqz at key,
-  replace key := mul_le_mul_of_nonneg_left key h₃.le,
-  field_simp [h₁.ne', h₂.ne', h₃.ne', mul_comm (z - x) _] at key ⊢,
-  rw div_le_div_right,
-  { linarith },
-  { nlinarith }
-end
-
-/-- For a function `f` defined on a convex subset `D` of `𝕜`, `f` is convex on `D` iff, for any
-three points `x < y < z` the slope of the secant line of `f` on `[x, y]` is less than or equal to
-the slope,of the secant line of `f` on `[x, z]`. -/
-lemma convex_on_iff_slope_mono_adjacent {s : set 𝕜} (hs : convex 𝕜 s) {f : 𝕜 → 𝕜} :
-  convex_on 𝕜 s f ↔
-  (∀ {x y z : 𝕜}, x ∈ s → z ∈ s → x < y → y < z →
-    (f y - f x) / (y - x) ≤ (f z - f y) / (z - y)) :=
-⟨convex_on.slope_mono_adjacent, convex_on_of_slope_mono_adjacent hs⟩
-
-/-- For a function `f` defined on a convex subset `D` of `𝕜`, if for any three points `x < y < z`
-the slope of the secant line of `f` on `[x, y]` is greater than or equal to the slope
-of the secant line of `f` on `[x, z]`, then `f` is concave on `D`. -/
-lemma concave_on_of_slope_mono_adjacent {s : set 𝕜} (hs : convex 𝕜 s) {f : 𝕜 → 𝕜}
-  (hf : ∀ {x y z : 𝕜}, x ∈ s → z ∈ s → x < y → y < z →
-    (f z - f y) / (z - y) ≤ (f y - f x) / (y - x)) : concave_on 𝕜 s f :=
-begin
-  rw ←neg_convex_on_iff,
-  refine convex_on_of_slope_mono_adjacent hs (λ x y z hx hz hxy hyz, _),
-  rw ←neg_le_neg_iff,
-  simp_rw [←neg_div, neg_sub, pi.neg_apply, neg_sub_neg],
-  exact hf hx hz hxy hyz,
-end
-
-/-- For a function `f` defined on a subset `D` of `𝕜`, if `f` is concave on `D`, then for any three
-points `x < y < z`, the slope of the secant line of `f` on `[x, y]` is greater than or equal to the
-slope of the secant line of `f` on `[x, z]`. -/
-lemma concave_on.slope_mono_adjacent {s : set 𝕜} {f : 𝕜 → 𝕜} (hf : concave_on 𝕜 s f)
-  {x y z : 𝕜} (hx : x ∈ s) (hz : z ∈ s) (hxy : x < y) (hyz : y < z) :
-  (f z - f y) / (z - y) ≤ (f y - f x) / (y - x) :=
-begin
-  rw [←neg_le_neg_iff, ←neg_sub_neg (f x), ←neg_sub_neg (f y)],
-  simp_rw [←pi.neg_apply, ←neg_div, neg_sub],
-  exact convex_on.slope_mono_adjacent hf.neg hx hz hxy hyz,
-end
-
-/-- For a function `f` defined on a convex subset `D` of `𝕜`, `f` is concave on `D` iff for any
-three points `x < y < z` the slope of the secant line of `f` on `[x, y]` is greater than or equal to
-the slope of the secant line of `f` on `[x, z]`. -/
-lemma concave_on_iff_slope_mono_adjacent {s : set 𝕜} (hs : convex 𝕜 s) {f : 𝕜 → 𝕜} :
-  concave_on 𝕜 s f ↔
-  (∀ {x y z : 𝕜}, x ∈ s → z ∈ s → x < y → y < z →
-    (f z - f y) / (z - y) ≤ (f y - f x) / (y - x)) :=
-⟨concave_on.slope_mono_adjacent, concave_on_of_slope_mono_adjacent hs⟩
-
 end has_scalar
 end ordered_add_comm_monoid
 end linear_ordered_field
diff --git a/src/analysis/convex/slope.lean b/src/analysis/convex/slope.lean
new file mode 100644
index 0000000000000..4fac058af33e9
--- /dev/null
+++ b/src/analysis/convex/slope.lean
@@ -0,0 +1,113 @@
+/-
+Copyright (c) 2021 Yury Kudriashov. All rights reserved.
+Released under Apache 2.0 license as described in the file LICENSE.
+Authors: Yury Kudriashov, Malo Jaffré
+-/
+import analysis.convex.function
+
+/-!
+# Slopes of convex functions
+
+This file relates convexity/concavity of functions in a linearly ordered field and the monotonicity
+of their slopes.
+
+The main use is to show convexity/concavity from monotonicity of the derivative.
+-/
+
+variables {𝕜 : Type*} [linear_ordered_field 𝕜] {s : set 𝕜} {f : 𝕜 → 𝕜}
+
+/-- If `f : 𝕜 → 𝕜` is convex, then for any three points `x < y < z` the slope of the secant line of
+`f` on `[x, y]` is less than the slope of the secant line of `f` on `[x, z]`. -/
+lemma convex_on.slope_mono_adjacent (hf : convex_on 𝕜 s f)
+  {x y z : 𝕜} (hx : x ∈ s) (hz : z ∈ s) (hxy : x < y) (hyz : y < z) :
+  (f y - f x) / (y - x) ≤ (f z - f y) / (z - y) :=
+begin
+  have hxz := hxy.trans hyz,
+  rw ←sub_pos at hxy hxz hyz,
+  suffices : f y / (y - x) + f y / (z - y) ≤ f x / (y - x) + f z / (z - y),
+  { ring_nf at this ⊢, linarith },
+  set a := (z - y) / (z - x),
+  set b := (y - x) / (z - x),
+  have hy : a • x + b • z = y, by { field_simp, rw div_eq_iff; [ring, linarith] },
+  have key, from
+    hf.2 hx hz
+      (show 0 ≤ a, by apply div_nonneg; linarith)
+      (show 0 ≤ b, by apply div_nonneg; linarith)
+      (show a + b = 1, by { field_simp, rw div_eq_iff; [ring, linarith] }),
+  rw hy at key,
+  replace key := mul_le_mul_of_nonneg_left key hxz.le,
+  field_simp [hxy.ne', hyz.ne', hxz.ne', mul_comm (z - x) _] at key ⊢,
+  rw div_le_div_right,
+  { linarith },
+  { nlinarith }
+end
+
+/-- If `f : 𝕜 → 𝕜` is concave, then for any three points `x < y < z` the slope of the secant line of
+`f` on `[x, y]` is greater than the slope of the secant line of `f` on `[x, z]`. -/
+lemma concave_on.slope_anti_adjacent (hf : concave_on 𝕜 s f) {x y z : 𝕜} (hx : x ∈ s)
+  (hz : z ∈ s) (hxy : x < y) (hyz : y < z) :
+  (f z - f y) / (z - y) ≤ (f y - f x) / (y - x) :=
+begin
+  rw [←neg_le_neg_iff, ←neg_sub_neg (f x), ←neg_sub_neg (f y)],
+  simp_rw [←pi.neg_apply, ←neg_div, neg_sub],
+  exact convex_on.slope_mono_adjacent hf.neg hx hz hxy hyz,
+end
+
+/-- If for any three points `x < y < z`, the slope of the secant line of `f : 𝕜 → 𝕜` on `[x, y]` is
+less than the slope of the secant line of `f` on `[x, z]`, then `f` is convex. -/
+lemma convex_on_of_slope_mono_adjacent (hs : convex 𝕜 s)
+  (hf : ∀ {x y z : 𝕜}, x ∈ s → z ∈ s → x < y → y < z →
+    (f y - f x) / (y - x) ≤ (f z - f y) / (z - y)) :
+  convex_on 𝕜 s f :=
+linear_order.convex_on_of_lt hs
+begin
+  assume x z hx hz hxz a b ha hb hab,
+  let y := a * x + b * z,
+  have hxy : x < y,
+  { rw [← one_mul x, ← hab, add_mul],
+    exact add_lt_add_left ((mul_lt_mul_left hb).2 hxz) _ },
+  have hyz : y < z,
+  { rw [← one_mul z, ← hab, add_mul],
+    exact add_lt_add_right ((mul_lt_mul_left ha).2 hxz) _ },
+  have : (f y - f x) * (z - y) ≤ (f z - f y) * (y - x),
+    from (div_le_div_iff (sub_pos.2 hxy) (sub_pos.2 hyz)).1 (hf hx hz hxy hyz),
+  have hxz : 0 < z - x, from sub_pos.2 (hxy.trans hyz),
+  have ha : (z - y) / (z - x) = a,
+  { rw [eq_comm, ← sub_eq_iff_eq_add'] at hab,
+    simp_rw [div_eq_iff hxz.ne', y, ←hab], ring },
+  have hb : (y - x) / (z - x) = b,
+  { rw [eq_comm, ← sub_eq_iff_eq_add] at hab,
+    simp_rw [div_eq_iff hxz.ne', y, ←hab], ring },
+  rwa [sub_mul, sub_mul, sub_le_iff_le_add', ← add_sub_assoc, le_sub_iff_add_le, ← mul_add,
+    sub_add_sub_cancel, ← le_div_iff hxz, add_div, mul_div_assoc, mul_div_assoc, mul_comm (f x),
+    mul_comm (f z), ha, hb] at this,
+end
+
+/-- If for any three points `x < y < z`, the slope of the secant line of `f : 𝕜 → 𝕜` on `[x, y]` is
+greater than the slope of the secant line of `f` on `[x, z]`, then `f` is concave. -/
+lemma concave_on_of_slope_anti_adjacent (hs : convex 𝕜 s)
+  (hf : ∀ {x y z : 𝕜}, x ∈ s → z ∈ s → x < y → y < z →
+    (f z - f y) / (z - y) ≤ (f y - f x) / (y - x)) : concave_on 𝕜 s f :=
+begin
+  rw ←neg_convex_on_iff,
+  refine convex_on_of_slope_mono_adjacent hs (λ x y z hx hz hxy hyz, _),
+  rw ←neg_le_neg_iff,
+  simp_rw [←neg_div, neg_sub, pi.neg_apply, neg_sub_neg],
+  exact hf hx hz hxy hyz,
+end
+
+/-- A function `f : 𝕜 → 𝕜` is convex iff for any three points `x < y < z` the slope of the secant
+line of `f` on `[x, y]` is less than the slope of the secant line of `f` on `[x, z]`. -/
+lemma convex_on_iff_slope_mono_adjacent :
+  convex_on 𝕜 s f ↔ convex 𝕜 s ∧
+    ∀ ⦃x y z : 𝕜⦄, x ∈ s → z ∈ s → x < y → y < z →
+      (f y - f x) / (y - x) ≤ (f z - f y) / (z - y) :=
+⟨λ h, ⟨h.1, λ x y z, h.slope_mono_adjacent⟩, λ h, convex_on_of_slope_mono_adjacent h.1 h.2⟩
+
+/-- A function `f : 𝕜 → 𝕜` is concave iff for any three points `x < y < z` the slope of the secant
+line of `f` on `[x, y]` is greater than the slope of the secant line of `f` on `[x, z]`. -/
+lemma concave_on_iff_slope_anti_adjacent :
+  concave_on 𝕜 s f ↔ convex 𝕜 s ∧
+    ∀ ⦃x y z : 𝕜⦄, x ∈ s → z ∈ s → x < y → y < z →
+      (f z - f y) / (z - y) ≤ (f y - f x) / (y - x) :=
+⟨λ h, ⟨h.1, λ x y z, h.slope_anti_adjacent⟩, λ h, concave_on_of_slope_anti_adjacent h.1 h.2⟩