forked from stacks/stacks-project
-
Notifications
You must be signed in to change notification settings - Fork 0
/
restricted.tex
7376 lines (6594 loc) · 277 KB
/
restricted.tex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
\input{preamble}
% OK, start here.
%
\begin{document}
\title{Algebraization of Formal Spaces}
\maketitle
\phantomsection
\label{section-phantom}
\tableofcontents
\section{Introduction}
\label{section-introduction}
\noindent
The main goal of this chapter is to prove Artin's theorem on dilatations,
see Theorem \ref{theorem-dilatations}; the result on contractions will be
discussed in Artin's Axioms, Section \ref{artin-section-contractions}.
Both results use some material on formal algebraic spaces, hence
in the middle part of this chapter, we
continue the discussion of formal algebraic spaces from the previous
chapter, see Formal Spaces, Section \ref{formal-spaces-section-introduction}.
The first part of this chapter is dedicated to algebraic preliminaries,
mostly dealing with algebraization of rig-\'etale algebras.
\medskip\noindent
Let $A$ be a Noetherian ring and let $I \subset A$ be an ideal. In the
first part of this chapter (Sections \ref{section-two-categories}
-- \ref{section-approximation-principal})
we discuss the category of $I$-adically complete algebras $B$
topologically of finite type over a Noetherian ring $A$.
It is shown that $B = A\{x_1, \ldots, x_n\}/J$ for some
(closed) ideal $J$ in the restricted power series ring
(where $A$ is endowed with the $I$-adic topology).
We show there is a good notion of a naive cotangent complex
$\NL_{B/A}^\wedge$. If some power of $I$ annihilates
$\NL_{B/A}^\wedge$, then we say $B$ is a rig-\'etale algebra
over $(A, I)$; there is a similar notion of rig-smooth algebras.
If $A$ is a G-ring, then we can show, using Popescu's theorem,
that any rig-smooth algebra $B$ over $(A, I)$ is the completion
of a finite type $A$-algebra; informally we say that we can ``algebraize'' $B$.
However, the main result of the first part is that any rig-\'etale
algebra $B$ over $(A, I)$ can be algebraized, see
Lemma \ref{lemma-approximate}.
One thing to note here is that we prove this without assuming
the ring $A$ is a G-ring.
\medskip\noindent
Many of the results discussed in the first part can be found in the paper
\cite{Elkik}. Other general references for this part are
\cite{EGA}, \cite{Abbes}, and \cite{Fujiwara-Kato}.
\medskip\noindent
In the second part of this chapter
(Sections \ref{section-finite-type-red} -- \ref{section-formal-modifications})
we talk about types of morphisms of formal algebraic spaces
in a reasonable level of generality (mostly for locally Noetherian
formal algebraic spaces). The most interesting of these is
the notion of a ``formal modification'' in the last section.
We carefully check that our definition agrees with
Artin's definition in \cite{ArtinII}.
\medskip\noindent
Finally, in the third and last part of this chapter
(Sections \ref{section-completion-and-morphisms} --
\ref{section-modifications}) we prove the main theorem and
we give a few applications.
In fact, we deduce Artin's theorem from a stronger
result, namely, Theorem \ref{theorem-dilatations-general}.
This theorem says very roughly: if $f : \mathfrak X \to \mathfrak X'$
is a rig-\'etale morphism and $\mathfrak X'$ is the formal
completion of a locally Noetherian algebraic space, then so
is $\mathfrak X$. In Artin's work the morphism $f$ is assumed proper
and rig-surjective.
\section{Two categories}
\label{section-two-categories}
\noindent
Let $A$ be a ring and let $I \subset A$ be an ideal.
In this section ${}^\wedge$ will mean $I$-adic completion.
Set $A_n = A/I^n$ so that the $I$-adic completion of $A$ is
$A^\wedge = \lim A_n$. Let $\mathcal{C}$ be the
category
\begin{equation}
\label{equation-C}
\mathcal{C} =
\left\{
\begin{matrix}
\text{inverse systems }\ldots \to B_3 \to B_2 \to B_1 \\
\text{where }B_n\text{ is a finite type }A_n\text{-algebra,}\\
B_{n + 1} \to B_n\text{ is an }A_{n + 1}\text{-algebra map}\\
\text{which induces }B_{n + 1}/I^nB_{n + 1} \cong B_n
\end{matrix}
\right\}
\end{equation}
Morphisms in $\mathcal{C}$ are given by systems of homomorphisms.
Let $\mathcal{C}'$ be the category
\begin{equation}
\label{equation-C-prime}
\mathcal{C}' =
\left\{
\begin{matrix}
A\text{-algebras }B\text{ which are }I\text{-adically complete}\\
\text{such that }B/IB\text{ is of finite type over }A/I
\end{matrix}
\right\}
\end{equation}
Morphisms in $\mathcal{C}'$ are $A$-algebra maps. There is a functor
\begin{equation}
\label{equation-from-complete-to-systems}
\mathcal{C}' \longrightarrow \mathcal{C},\quad
B \longmapsto (B/I^nB)
\end{equation}
Indeed, since $B/IB$ is of finite type over $A/I$ the ring maps
$A_n = A/I^n \to B/I^nB$ are of finite type by
Algebra, Lemma \ref{algebra-lemma-finite-type-mod-nilpotent}.
\begin{lemma}
\label{lemma-topologically-finite-type}
Let $A$ be a ring and let $I \subset A$ be a finitely generated ideal.
The functor
$$
\mathcal{C} \longrightarrow \mathcal{C}',\quad
(B_n) \longmapsto B = \lim B_n
$$
is a quasi-inverse to (\ref{equation-from-complete-to-systems}).
The completions $A[x_1, \ldots, x_r]^\wedge$ are in $\mathcal{C}'$ and
any object of $\mathcal{C}'$ is of the form
$$
B = A[x_1, \ldots, x_r]^\wedge / J
$$
for some ideal $J \subset A[x_1, \ldots, x_r]^\wedge$.
\end{lemma}
\begin{proof}
Let $(B_n)$ be an object of $\mathcal{C}$. By
Algebra, Lemma \ref{algebra-lemma-limit-complete}
we see that $B = \lim B_n$ is $I$-adically complete
and $B/I^nB = B_n$. Hence we see that $B$ is an object of
$\mathcal{C}'$ and that we can recover the object $(B_n)$
by taking the quotients.
Conversely, if $B$ is an object of $\mathcal{C}'$, then
$B = \lim B/I^nB$ by assumption. Thus $B \mapsto (B/I^nB)$ is a quasi-inverse
to the functor of the lemma.
\medskip\noindent
Since $A[x_1, \ldots, x_r]^\wedge = \lim A_n[x_1, \ldots, x_r]$
it is an object of $\mathcal{C}'$ by the first statement of the lemma.
Finally, let $B$ be an object of $\mathcal{C}'$. Choose
$b_1, \ldots, b_r \in B$ whose images in $B/IB$ generate
$B/IB$ as an algebra over $A/I$. Since $B$ is $I$-adically
complete, the $A$-algebra map $A[x_1, \ldots, x_r] \to B$, $x_i \mapsto b_i$
extends to an $A$-algebra map $A[x_1, \ldots, x_r]^\wedge \to B$.
To finish the proof we have to show this map is surjective
which follows from Algebra, Lemma \ref{algebra-lemma-completion-generalities}
as our map $A[x_1, \ldots, x_r] \to B$ is surjective modulo $I$
and as $B = B^\wedge$.
\end{proof}
\noindent
We warn the reader that, in case $A$ is not Noetherian, the
quotient of an object of $\mathcal{C}'$ may not be an object
of $\mathcal{C}'$. See Examples, Lemma
\ref{examples-lemma-noncomplete-quotient}.
Next we show this does not happen when $A$ is Noetherian.
\begin{lemma}
\label{lemma-topologically-finite-type-Noetherian}
\begin{reference}
\cite[Proposition 7.5.5]{EGA1}
\end{reference}
Let $A$ be a Noetherian ring and let $I \subset A$ be an ideal. Then
\begin{enumerate}
\item every object of the category $\mathcal{C}'$
(\ref{equation-C-prime}) is Noetherian,
\item if $B \in \Ob(\mathcal{C}')$ and $J \subset B$ is an ideal,
then $B/J$ is an object of $\mathcal{C}'$,
\item for a finite type $A$-algebra $C$ the $I$-adic completion
$C^\wedge$ is in $\mathcal{C}'$,
\item in particular the completion $A[x_1, \ldots, x_r]^\wedge$
is in $\mathcal{C}'$.
\end{enumerate}
\end{lemma}
\begin{proof}
Part (4) follows from
Algebra, Lemma \ref{algebra-lemma-completion-Noetherian-Noetherian}
as $A[x_1, \ldots, x_r]$ is Noetherian
(Algebra, Lemma \ref{algebra-lemma-Noetherian-permanence}).
To see (1) by Lemma \ref{lemma-topologically-finite-type}
we reduce to the case of the completion of the polynomial ring
which we just proved.
Part (2) follows from Algebra, Lemma \ref{algebra-lemma-completion-tensor}
which tells us that ever finite $B$-module is $IB$-adically complete.
Part (3) follows in the same manner as part (4).
\end{proof}
\begin{remark}[Base change]
\label{remark-base-change}
Let $\varphi : A_1 \to A_2$ be a ring map and let
$I_i \subset A_i$ be ideals such that $\varphi(I_1^c) \subset I_2$
for some $c \geq 1$. This induces ring maps
$A_{1, cn} = A_1/I_1^{cn} \to A_2/I_2^n = A_{2, n}$ for all $n \geq 1$.
Let $\mathcal{C}_i$ be the category (\ref{equation-C}) for $(A_i, I_i)$.
There is a base change functor
\begin{equation}
\label{equation-base-change-systems}
\mathcal{C}_1 \longrightarrow \mathcal{C}_2,\quad
(B_n) \longmapsto (B_{cn} \otimes_{A_{1, cn}} A_{2, n})
\end{equation}
Let $\mathcal{C}_i'$ be the category (\ref{equation-C-prime}) for $(A_i, I_i)$.
If $I_2$ is finitely generated, then there is a base change functor
\begin{equation}
\label{equation-base-change-complete}
\mathcal{C}_1' \longrightarrow \mathcal{C}_2',\quad
B \longmapsto (B \otimes_{A_1} A_2)^\wedge
\end{equation}
because in this case the completion is complete
(Algebra, Lemma \ref{algebra-lemma-hathat-finitely-generated}).
If both $I_1$ and $I_2$ are finitely generated, then
the two base change functors agree via the functors
(\ref{equation-from-complete-to-systems})
which are equivalences by Lemma \ref{lemma-topologically-finite-type}.
\end{remark}
\begin{remark}[Base change by closed immersion]
\label{remark-take-bar}
Let $A$ be a Noetherian ring and $I \subset A$ an ideal.
Let $\mathfrak a \subset A$ be an ideal. Denote $\bar A = A/\mathfrak a$.
Let $\bar I \subset \bar A$ be an ideal such that
$I^c \bar A \subset \bar I$ and $\bar I^d \subset I\bar A$
for some $c, d \geq 1$. In this case the base change functor
(\ref{equation-base-change-complete}) for $(A, I)$ to $(\bar A, \bar I)$
is given by $B \mapsto \bar B = B/\mathfrak aB$. Namely, we have
\begin{equation}
\label{equation-base-change-to-closed}
\bar B = (B \otimes_A \bar A)^\wedge = (B/\mathfrak a B)^\wedge =
B/\mathfrak a B
\end{equation}
the last equality because any finite $B$-module is $I$-adically complete by
Algebra, Lemma \ref{algebra-lemma-completion-tensor}
and if annihilated by $\mathfrak a$ also $\bar I$-adically complete by
Algebra, Lemma \ref{algebra-lemma-change-ideal-completion}.
\end{remark}
\section{A naive cotangent complex}
\label{section-naive-cotangent-complex}
\noindent
Let $A$ be a Noetherian ring and let $I \subset A$ be a ideal.
Let $B$ be an $A$-algebra which is $I$-adically complete such
that $A/I \to B/IB$ is of finite type, i.e., an object of
(\ref{equation-C-prime}).
By Lemma \ref{lemma-topologically-finite-type-Noetherian} we can write
$$
B = A[x_1, \ldots, x_r]^\wedge / J
$$
for some finitely generated ideal $J$. For a choice of presentation as
above we define the {\it naive cotangent complex}
in this setting by the formula
\begin{equation}
\label{equation-NL}
\NL_{B/A}^\wedge = (J/J^2 \longrightarrow \bigoplus B\text{d}x_i)
\end{equation}
with terms sitting in degrees $-1$ and $0$
where the map sends the residue class of $g \in J$ to the differential
$\text{d}g = \sum (\partial g/\partial x_i) \text{d}x_i$. Here
the partial derivative is taken by thinking of $g$ as a power series.
The following lemma shows that $\NL_{B/A}^\wedge$ is well defined
up to homotopy.
\begin{lemma}
\label{lemma-NL-up-to-homotopy}
Let $A$ be a Noetherian ring and let $I \subset A$ be a ideal.
Let $B$ be an object of (\ref{equation-C-prime}). The naive
cotangent complex $\NL_{B/A}^\wedge$ is well defined in $K(B)$.
\end{lemma}
\begin{proof}
The lemma signifies that given a second presentation
$B = A[y_1, \ldots, y_s]^\wedge / K$ the complexes of $B$-modules
$$
(J/J^2 \to B\text{d}x_i)
\quad\text{and}\quad
(K/K^2 \to \bigoplus B\text{d}y_j)
$$
are homotopy equivalent. To see this, we can argue exactly as in
the proof of Algebra, Lemma \ref{algebra-lemma-NL-homotopy}.
\medskip\noindent
Step 1. If we choose $g_i(y_1, \ldots, y_s) \in A[y_1, \ldots, y_s]^\wedge$
mapping to the image of $x_i$ in $B$, then we obtain a (unique) continuous
$A$-algebra homomorphism
$$
A[x_1, \ldots, x_r]^\wedge \to A[y_1, \ldots, y_s]^\wedge,\quad
x_i \mapsto g_i(y_1, \ldots, y_s)
$$
compatible with the given surjections to $B$. Such a map is called
a morphism of presentations. It induces a map from $J$ into $K$
and hence induces a $B$-module map $J/J^2 \to K/K^2$. Sending
$\text{d}x_i$ to $\sum (\partial g_i/\partial y_j)\text{d}y_j$
we obtain a map of complexes
$$
(J/J^2 \to \bigoplus B\text{d}x_i)
\longrightarrow
(K/K^2 \to \bigoplus B\text{d}y_j)
$$
Of course we can do the same thing with the roles of the two presentations
exchanged to get a map of complexes in the other direction.
\medskip\noindent
Step 2. The construction above is compatible with compositions of morphsms
of presentations. Hence to finish the proof it suffices to show: given
$g_i(x_1, \ldots, x_r) \in A[x_1, \ldots, x_n]^\wedge$
mapping to the image of $x_i$ in $B$, the induced map of complexes
$$
(J/J^2 \to \bigoplus B\text{d}x_i)
\longrightarrow
(J/J^2 \to \bigoplus B\text{d}x_i)
$$
is homotopic to the identity map. To see this consider the map
$h : \bigoplus B \text{d}x_i \to J/J^2$ given by the rule
$\text{d}x_i \mapsto g_i(x_1, \ldots, x_n) - x_i$ and compute.
\end{proof}
\begin{lemma}
\label{lemma-NL-is-completion}
Let $A$ be a Noetherian ring and let $I \subset A$ be a ideal.
Let $A \to B$ be a finite type ring map. Choose a presentation
$\alpha : A[x_1, \ldots, x_n] \to B$. Then
$\NL_{B^\wedge/A}^\wedge = \lim \NL(\alpha) \otimes_B B^\wedge$
as complexes and
$\NL_{B^\wedge/A}^\wedge = \NL_{B/A} \otimes_B^\mathbf{L} B^\wedge$
in $D(B^\wedge)$.
\end{lemma}
\begin{proof}
The statement makes sense as $B^\wedge$ is an object of
(\ref{equation-C-prime}) by
Lemma \ref{lemma-topologically-finite-type-Noetherian}.
Let $J = \Ker(\alpha)$. The functor of taking
$I$-adic completion is exact on finite modules over
$A[x_1, \ldots, x_n]$ and agrees with the functor
$M \mapsto M \otimes_{A[x_1, \ldots, x_n]} A[x_1, \ldots, x_n]^\wedge$, see
Algebra, Lemmas \ref{algebra-lemma-completion-tensor} and
\ref{algebra-lemma-completion-flat}. Moreover, the ring maps
$A[x_1, \ldots, x_n] \to A[x_1, \ldots, x_n]^\wedge$
and $B \to B^\wedge$ are flat. Hence
$B^\wedge = A[x_1, \ldots, x_n]^\wedge / J^\wedge$ and
$$
(J/J^2) \otimes_B B^\wedge = (J/J^2)^\wedge = J^\wedge/(J^\wedge)^2
$$
Since $\NL(\alpha) = (J/J^2 \to \bigoplus B\text{d}x_i)$,
see Algebra, Section \ref{algebra-section-netherlander},
we conclude the complex $\NL_{B^\wedge/A}^\wedge$ is equal
to $\NL(\alpha) \otimes_B B^\wedge$. The final statement
follows as $\NL_{B/A}$ is homotopy equivalent to $\NL(\alpha)$
and because the ring map $B \to B^\wedge$ is flat (so derived
base change along $B \to B^\wedge$ is just base change).
\end{proof}
\begin{lemma}
\label{lemma-NL-is-limit}
Let $A$ be a Noetherian ring and let $I \subset A$ be a ideal.
Let $B$ be an object of (\ref{equation-C-prime}). Then
\begin{enumerate}
\item the pro-objects
$\{\NL_{B/A}^\wedge \otimes_B B/I^nB\}$ and $\{\NL_{B_n/A_n}\}$
of $D(B)$ are strictly isomorphic (see proof for elucidation),
\item $\NL_{B/A}^\wedge = R\lim \NL_{B_n/A_n}$ in $D(B)$.
\end{enumerate}
Here $B_n$ and $A_n$ are as in Section \ref{section-two-categories}.
\end{lemma}
\begin{proof}
The statement means the following: for every $n$ we have a well
defined complex $\NL_{B_n/A_n}$ of $B_n$-modules and we have
transition maps $\NL_{B_{n + 1}/A_{n + 1}} \to \NL_{B_n/A_n}$.
See Algebra, Section \ref{algebra-section-netherlander}.
Thus we can consider
$$
\ldots \to \NL_{B_3/A_3} \to \NL_{B_2/A_2} \to \NL_{B_1/A_1}
$$
as an inverse system of complexes of $B$-modules and a fortiori as
an inverse system in $D(B)$. Furthermore $R\lim \NL_{B_n/A_n}$
is a homotopy limit of this inverse system, see
Derived Categories, Section \ref{derived-section-derived-limit}.
\medskip\noindent
Choose a presentation $B = A[x_1, \ldots, x_r]^\wedge / J$. This
defines presentations
$$
B_n = B/I^nB = A_n[x_1, \ldots, x_r]/J_n
$$
where
$$
J_n = JA_n[x_1, \ldots, x_r] =
J/(J \cap I^nA[x_1, \ldots, x_r]^\wedge)
$$
The two term complex $J_n/J_n^2 \longrightarrow \bigoplus B_n \text{d}x_i$
represents $\NL_{B_n/A_n}$, see
Algebra, Section \ref{algebra-section-netherlander}.
By Artin-Rees (Algebra, Lemma \ref{algebra-lemma-Artin-Rees})
in the Noetherian ring $A[x_1, \ldots, x_r]^\wedge$
(Lemma \ref{lemma-topologically-finite-type-Noetherian})
we find a $c \geq 0$ such that we have canonical surjections
$$
J/I^nJ \to J_n \to J/I^{n - c}J \to J_{n - c},\quad n \geq c
$$
for all $n \geq c$. A moment's thought shows that these maps are
compatible with differentials and we obtain maps of complexes
$$
\NL_{B/A}^\wedge \otimes_B B/I^nB \to
\NL_{B_n/A_n} \to
\NL_{B/A}^\wedge \otimes_B B/I^{n - c}B \to
\NL_{B_{n - c}/A_{n - c}}
$$
compatible with the transition maps of the inverse systems
$\{\NL_{B/A}^\wedge \otimes_B B/I^nB\}$ and $\{\NL_{B_n/A_n}\}$.
This proves part (1) of the lemma.
\medskip\noindent
By part (1) and since pro-isomorphic
systems have the same $R\lim$ in order to prove (2)
it suffices to show that $\NL_{B/A}^\wedge$ is
equal to $R\lim \NL_{B/A}^\wedge \otimes_B B/I^nB$.
However, $\NL_{B/A}^\wedge$ is a two term complex $M^\bullet$
of finite $B$-modules which are $I$-adically complete for example by
Algebra, Lemma \ref{algebra-lemma-completion-tensor}. Hence
$M^\bullet = \lim M^\bullet/I^nM^\bullet = R\lim M^\bullet/I^n M^\bullet$, see
More on Algebra, Lemma \ref{more-algebra-lemma-compute-Rlim-modules} and
Remark \ref{more-algebra-remark-how-unique}.
\end{proof}
\begin{lemma}
\label{lemma-NL-base-change}
Let $(A_1, I_1) \to (A_2, I_2)$ be as in
Remark \ref{remark-base-change} with $A_1$ and $A_2$ Noetherian.
Let $B_1$ be in (\ref{equation-C-prime}) for $(A_1, I_1)$.
Let $B_2$ be the base change of $B_1$. Then there is a canonical map
$$
\NL_{B_1/A_1} \otimes_{B_2} B_1 \to \NL_{B_2/A_2}
$$
which induces and isomorphism on $H^0$ and a surjection on $H^{-1}$.
\end{lemma}
\begin{proof}
Choose a presentation $B_1 = A_1[x_1, \ldots, x_r]^\wedge/J_1$.
Since
$A_2/I_2^n[x_1, \ldots, x_r] =
A_1/I_1^{cn}[x_1, \ldots, x_r] \otimes_{A_1/I_1^{cn}} A_2/I_2^n$
we have
$$
A_2[x_1, \ldots, x_r]^\wedge =
(A_1[x_1, \ldots, x_r]^\wedge \otimes_{A_1} A_2)^\wedge
$$
where we use $I_2$-adic completion on both sides (but of course
$I_1$-adic completion for $A_1[x_1, \ldots, x_r]^\wedge$).
Set $J_2 = J_1 A_2[x_1, \ldots, x_r]^\wedge$. Arguing similarly
we get the presentation
\begin{align*}
B_2
& =
(B_1 \otimes_{A_1} A_2)^\wedge \\
& =
\lim \frac{A_1/I_1^{cn}[x_1, \ldots, x_r]}{J_1(A_1/I_1^{cn}[x_1, \ldots, x_r])}
\otimes_{A_1/I_1^{cn}} A_2/I_2^n \\
& =
\lim \frac{A_2/I_2^n[x_1, \ldots, x_r]}{J_2(A_2/I_2^n[x_1, \ldots, x_r])} \\
& =
A_2[x_1, \ldots, x_r]^\wedge/J_2
\end{align*}
for $B_2$ over $A_2$. As a consequence obtain a commutative diagram
$$
\xymatrix{
\NL^\wedge_{B_1/A_1} : \ar[d] &
J_1/J_1^2 \ar[r]_-{\text{d}} \ar[d] & \bigoplus B_1\text{d}x_i \ar[d] \\
\NL^\wedge_{B_2/A_2} : &
J_2/J_2^2 \ar[r]^-{\text{d}} & \bigoplus B_2\text{d}x_i
}
$$
The induced arrow $J_1/J_1^2 \otimes_{B_1} B_2 \to J_2/J_2^2$
is surjective because $J_2$ is generated by the image of $J_1$.
This determines the arrow displayed in the lemma. We omit the proof
that this arrow is well defined up to homotopy (i.e., independent
of the choice of the presentations up to homotopy). The statement
about the induced map on cohomology modules follows easily
from the discussion (details omitted).
\end{proof}
\begin{lemma}
\label{lemma-exact-sequence-NL}
Let $A$ be a Noetherian ring and let $I \subset A$ be a ideal.
Let $B \to C$ be morphism of (\ref{equation-C-prime}). Then
there is an exact sequence
$$
\xymatrix{
C \otimes_B H^0(\NL_{B/A}^\wedge) \ar[r] &
H^0(\NL_{C/A}^\wedge) \ar[r] &
H^0(\NL_{C/B}^\wedge) \ar[r] & 0 \\
H^{-1}(\NL_{B/A}^\wedge \otimes_B C) \ar[r] &
H^{-1}(\NL_{C/A}^\wedge) \ar[r] &
H^{-1}(\NL_{C/B}^\wedge) \ar[llu]
}
$$
See proof for elucidation.
\end{lemma}
\begin{proof}
Observe that taking the tensor product $\NL_{B/A}^\wedge \otimes_B C$
makes sense as $\NL_{B/A}^\wedge$ is well defined up to homotopy by
Lemma \ref{lemma-NL-up-to-homotopy}.
Also, $(B, IB)$ is pair where $B$ is a Noetherian ring
(Lemma \ref{lemma-topologically-finite-type-Noetherian})
and $C$ is in the corresponding category (\ref{equation-C-prime}).
Thus all the terms in the $6$-term sequence are (well) defined.
\medskip\noindent
Choose a presentation $B = A[x_1, \ldots, x_r]^\wedge/J$.
Choose a presentation $C = B[y_1, \ldots, y_s]^\wedge/J'$.
Combinging these presentations gives a presentation
$$
C = A[x_1, \ldots, x_r, y_1, \ldots, y_s]^\wedge/K
$$
Then the reader verifies that we obtain a commutative diagram
$$
\xymatrix{
0 \ar[r] &
\bigoplus C \text{d}x_i \ar[r] &
\bigoplus C \text{d}x_i \oplus \bigoplus C \text{d}y_j \ar[r] &
\bigoplus C \text{d}y_j \ar[r] &
0 \\
&
J/J^2 \otimes_B C \ar[r] \ar[u] &
K/K^2 \ar[r] \ar[u] &
J'/(J')^2 \ar[r] \ar[u] &
0
}
$$
with exact rows. Note that the vertical arrow on the left hand side
is the tensor product of the arrow defining $\NL_{B/A}^\wedge$ with
$\text{id}_C$. The lemma follows by applying the snake lemma
(Algebra, Lemma \ref{algebra-lemma-snake}).
\end{proof}
\begin{lemma}
\label{lemma-transitive-lci-at-end}
With assumptions as in Lemma \ref{lemma-exact-sequence-NL}
assume that $B/I^nB \to C/I^nC$ is a local complete intersection
homomorphism for all $n$. Then
$H^{-1}(\NL_{B/A}^\wedge \otimes_B C) \to H^{-1}(\NL_{C/A}^\wedge)$
is injective.
\end{lemma}
\begin{proof}
For each $n \geq 1$ we set $A_n = A/I^n$, $B_n = B/I^nB$, and
$C_n = C/I^nC$. We have
\begin{align*}
H^{-1}(\NL_{B/A}^\wedge \otimes_B C)
& =
\lim H^{-1}(\NL_{B/A}^\wedge \otimes_B C_n) \\
& =
\lim H^{-1}(\NL_{B/A}^\wedge \otimes_B B_n \otimes_{B_n} C_n) \\
& =
\lim H^{-1}(\NL_{B_n/A_n} \otimes_{B_n} C_n)
\end{align*}
The first equality follows from
More on Algebra, Lemma \ref{more-algebra-lemma-consequence-Artin-Rees}
and the fact that $H^{-1}(\NL_{B/A}^\wedge \otimes_B C)$ is a finite
$C$-module and hence $I$-adically complete for example by
Algebra, Lemma \ref{algebra-lemma-completion-tensor}.
The second equality is trivial.
The third holds by Lemma \ref{lemma-NL-is-limit}.
The maps $H^{-1}(\NL_{B_n/A_n} \otimes_{B_n} C_n) \to
H^{-1}(\NL_{C_n/A_n})$ are injective by
More on Algebra, Lemma \ref{more-algebra-lemma-transitive-lci-at-end}.
The proof is finished because we also have
$H^{-1}(\NL_{C/A}^\wedge) = \lim H^{-1}(\NL_{C_n/A_n})$
similarly to the above.
\end{proof}
\section{Rig-smooth algebras}
\label{section-rig-smooth}
\noindent
As motivation for the following definition, please take a look at
More on Algebra, Remark \ref{more-algebra-remark-smoothness-ext-1-zero}.
\begin{definition}
\label{definition-rig-smooth-homomorphism}
Let $A$ be a Noetherian ring and let $I \subset A$ be an ideal.
Let $B$ be an object of (\ref{equation-C-prime}). We say
$B$ is {\it rig-smooth over $(A, I)$} if there exists an integer $c \geq 0$
such that $I^c$ annihilates $\Ext^1_B(\NL_{B/A}^\wedge, N)$ for every
$B$-module $N$.
\end{definition}
\noindent
Let us work out what this means.
\begin{lemma}
\label{lemma-equivalent-with-artin-smooth}
Let $A$ be a Noetherian ring and let $I \subset A$ be an ideal.
Let $B$ be an object of (\ref{equation-C-prime}). Write
$B = A[x_1, \ldots, x_r]^\wedge/J$
(Lemma \ref{lemma-topologically-finite-type-Noetherian})
and let $\NL_{B/A}^\wedge = (J/J^2 \to \bigoplus B\text{d}x_i)$
be its naive cotangent complex (\ref{equation-NL}).
The following are equivalent
\begin{enumerate}
\item $B$ is rig-smooth over $(A, I)$,
\item the object $\NL_{B/A}^\wedge$ of $D(B)$ satisfies the equivalent
conditions (1) -- (4) of More on Algebra, Lemma
\ref{more-algebra-lemma-ext-1-annihilated} with respect to the ideal $IB$,
\item there exists a $c \geq 0$ such that for all $a \in I^c$
there is a map $h : \bigoplus B\text{d}x_i \to J/J^2$ such that
$a : J/J^2 \to J/J^2$ is equal to $h \circ \text{d}$,
\item there exist $b_1, \ldots, b_s \in B$ such that
$V(b_1, \ldots, b_s) \subset V(IB)$ and such that for every
$l = 1, \ldots, s$ there exist $m \geq 0$, $f_1, \ldots, f_m \in J$,
and subset $T \subset \{1, \ldots, n\}$ with $|T| = m$ such that
\begin{enumerate}
\item $\det_{i \in T, j \leq m}(\partial f_j/ \partial x_i)$
divides $b_l$ in $B$, and
\item $b_l J \subset (f_1, \ldots, f_m) + J^2$.
\end{enumerate}
\end{enumerate}
\end{lemma}
\begin{proof}
The equivalence of (1), (2), and (3) is immediate from
More on Algebra, Lemma \ref{more-algebra-lemma-ext-1-annihilated}.
\medskip\noindent
Assume $b_1, \ldots, b_s$ are as in (4). Since $B$ is Noetherian the inclusion
$V(b_1, \ldots, b_s) \subset V(IB)$ implies $I^cB \subset (b_1, \ldots, b_s)$
for some $c \geq 0$ (for example by
Algebra, Lemma \ref{algebra-lemma-Noetherian-power-ideal-kills-module}).
Pick $1 \leq l \leq s$ and $m \geq 0$ and $f_1, \ldots, f_m \in J$
and $T \subset \{1, \ldots, n\}$ with $|T| = m$ satisfying (4)(a) and (b).
Then if we invert $b_l$ we see that
$$
\NL_{B/A}^\wedge \otimes_B B_{b_l} =
\left(
\bigoplus\nolimits_{j \leq m} B_{b_l} f_j
\longrightarrow
\bigoplus\nolimits_{i = 1, \ldots, n} B_{b_l} \text{d}x_i
\right)
$$
and moreover the arrow is isomorphic to the inclusion of the direct
summand $\bigoplus_{i \in T} B_{b_l} \text{d}x_i$. We conclude that
$H^{-1}(\NL_{B/A}^\wedge)$ is $b_l$-power torsion and that
$H^0(\NL_{B/A}^\wedge)$ becomes finite free after inverting $b_l$.
Combined with the inclusion $I^cB \subset (b_1, \ldots, b_s)$
we see that $H^{-1}(\NL_{B/A}^\wedge)$ is $IB$-power torsion.
Hence we see that condition (4) of
More on Algebra, Lemma \ref{more-algebra-lemma-ext-1-annihilated}
holds. In this way we see that (4) implies (2).
\medskip\noindent
Assume the equivalent conditions (1), (2), and (3) hold. We will prove
that (4) holds, but we strongly urge the reader to convince themselves
of this. The complex $\NL_{B/A}^\wedge$ determines an object of
$D^b_{\textit{Coh}}(\Spec(B))$ whose restriction to the Zariski open
$U = \Spec(B) \setminus V(IB)$ is a finite locally free module
$\mathcal{E}$ placed in degree $0$ (this follows for example from the
the fourth equivalent condition in
More on Algebra, Lemma \ref{more-algebra-lemma-ext-1-annihilated}).
Choose generators $f_1, \ldots, f_M$ for $J$.
This determines an exact sequence
$$
\bigoplus\nolimits_{j = 1, \ldots, M} \mathcal{O}_U \cdot f_j \to
\bigoplus\nolimits_{i = 1, \ldots, n} \mathcal{O}_U \cdot \text{d}x_i \to
\mathcal{E} \to 0
$$
Let $U = \bigcup_{l = 1, \ldots, s} U_l$
be a finite affine open covering such that
$\mathcal{E}|_{U_l}$ is free of rank $r_l = n - m_l$ for some integer
$n \geq m_l \geq 0$. After replacing
each $U_l$ by an affine open covering we may assume there exists
a subset $T_l \subset \{1, \ldots, n\}$ such that the elements
$\text{d}x_i$, $i \in \{1, \ldots, n\} \setminus T_l$ map to a
basis for $\mathcal{E}|_{U_l}$. Repeating the argument, we may
assume there exists a subset $T'_l \subset \{1, \ldots, M\}$
of cardinality $m_l$ such that $f_j$, $j \in T'_l$ map to a basis
of the kernel of $\mathcal{O}_{U_l} \cdot \text{d}x_i \to
\mathcal{E}|_{U_l}$. Finally, since the open covering
$U = \bigcup U_l$ may be refined by a open covering by standard opens
(Algebra, Lemma \ref{algebra-lemma-Zariski-topology})
we may assume $U_l = D(g_l)$ for some $g_l \in B$.
In particular we have $V(g_1, \ldots, g_s) = V(IB)$.
A linear algebra argument using our choices above shows that
$\det_{i \in T_l, j \in T'_l}(\partial f_j/ \partial x_i)$
maps to an invertible element of $B_{b_l}$. Similarly, the vanishing
of cohomology of $\NL_{B/A}^\wedge$ in degree $-1$ over $U_l$ shows that
$J/J^2 + (f_j; j \in T')$ is annihilated by a power of $b_l$.
After replacing each $g_l$ by a suitable power we obtain
conditions (4)(a) and (4)(b) of the lemma. Some details omitted.
\end{proof}
\begin{lemma}
\label{lemma-rig-smooth}
Let $A$ be a Noetherian ring and let $I$ be an ideal.
Let $B$ be a finite type $A$-algebra.
\begin{enumerate}
\item If $\Spec(B) \to \Spec(A)$ is smooth over $\Spec(A) \setminus V(I)$,
then $B^\wedge$ is rig-smooth over $(A, I)$.
\item If $B^\wedge$ is rig-smooth over $(A, I)$,
then there exists $g \in 1 + IB$ such that $\Spec(B_g)$ is smooth
over $\Spec(A) \setminus V(I)$.
\end{enumerate}
\end{lemma}
\begin{proof}
We will use Lemma \ref{lemma-equivalent-with-artin-smooth}
without further mention.
\medskip\noindent
Assume (1). Recall that formation of $\NL_{B/A}$ commutes with
localization, see Algebra, Lemma \ref{algebra-lemma-localize-NL}.
Hence by the very definition of smooth ring maps (in terms of
the naive cotangent complex being quasi-isomorphic to a finite
projective module placed in degree $0$), we see that
$\NL_{B/A}$ satisfies the fourth equivalent condition
of More on Algebra, Lemma \ref{more-algebra-lemma-ext-1-annihilated}
with respect to the ideal $IB$ (small detail omitted).
Since $\NL_{B^\wedge/A}^\wedge = \NL_{B/A} \otimes_B B^\wedge$
by Lemma \ref{lemma-NL-is-completion} we conclude (2) holds by
More on Algebra, Lemma
\ref{more-algebra-lemma-base-change-property-ext-1-annihilated}.
\medskip\noindent
Assume (2). Choose a presentation
$B = A[x_1, \ldots, x_n]/J$, set $N = J/J^2$, and
consider the element $\xi \in \Ext^1_B(\NL_{B/A}, J/J^2)$
determined by the identity map on $J/J^2$.
Using again that $\NL_{B^\wedge/A}^\wedge = \NL_{B/A} \otimes_B B^\wedge$
we find that our assumption implies the image
$$
\xi \otimes 1 \in
\Ext^1_{B^\wedge}(\NL_{B/A} \otimes_B B^\wedge, N \otimes_B B^\wedge) =
\Ext^1_{B^\wedge}(\NL_{B/A}, N) \otimes_B B^\wedge
$$
is annihilated by $I^c$ for some integer $c \geq 0$.
The equality holds for example by
More on Algebra, Lemma \ref{more-algebra-lemma-base-change-RHom}
(but can also easily be deduced from the much simpler
More on Algebra, Lemma
\ref{more-algebra-lemma-pseudo-coherence-and-base-change-ext}).
Thus $M = I^cB\xi \subset \Ext^1_B(\NL_{B/A}, N)$ is a finite submodule
which maps to zero in $\Ext^1_B(\NL_{B/A}, N) \otimes_B B^\wedge$.
Since $B \to B^\wedge$ is flat this means that
$M \otimes_B B^\wedge$ is zero. By
Nakayama's lemma (Algebra, Lemma \ref{algebra-lemma-NAK})
this means that $M = I^cB\xi$ is annihilated by an element
of the form $g = 1 + x$ with $x \in IB$.
This implies that for every $b \in I^cB$ there is a
$B$-linear dotted arrow making the diagram commute
$$
\xymatrix{
J/J^2 \ar[r] \ar[d]^b & \bigoplus B\text{d}x_i \ar@{..>}[d]^h \\
J/J^2 \ar[r] & (J/J^2)_g
}
$$
Thus $(\NL_{B/A})_{gb}$ is quasi-isomorphic
to a finite projective module; small detail omitted.
Since $(\NL_{B/A})_{gb} = \NL_{B_{gb}/A}$ in $D(B_{gb})$
this shows that $B_{gb}$ is smooth over $\Spec(A)$.
As this holds for all $b \in I^cB$ we conclude
that $\Spec(B_g) \to \Spec(A)$ is smooth over $\Spec(A) \setminus V(I)$
as desired.
\end{proof}
\begin{lemma}
\label{lemma-zero-ext-1-after-modding-out}
Let $(A_1, I_1) \to (A_2, I_2)$ be as in
Remark \ref{remark-base-change} with $A_1$ and $A_2$ Noetherian.
Let $B_1$ be in (\ref{equation-C-prime}) for $(A_1, I_1)$.
Let $B_2$ be the base change of $B_1$. Let $f_1 \in B_1$
with image $f_2 \in B_2$.
If $\Ext^1_{B_1}(\NL_{B_1/A_1}^\wedge, N_1)$ is annihilated
by $f_1$ for every $B_1$-module $N_1$, then
$\Ext^1_{B_2}(\NL_{B_2/A_2}^\wedge, N_2)$ is annihilated
by $f_2$ for every $B_2$-module $N_2$.
\end{lemma}
\begin{proof}
By Lemma \ref{lemma-NL-base-change} there is a map
$$
\NL_{B_1/A_1} \otimes_{B_2} B_1 \to \NL_{B_2/A_2}
$$
which induces and isomorphism on $H^0$ and a surjection on $H^{-1}$.
Thus the result by More on Algebra, Lemmas
\ref{more-algebra-lemma-two-term-base-change},
\ref{more-algebra-lemma-base-change-property-ext-1-annihilated}, and
\ref{more-algebra-lemma-surjection-property-ext-1-annihilated}
the last two applied with the principal ideals $(f_1) \subset B_1$ and
$(f_2) \subset B_2$.
\end{proof}
\begin{lemma}
\label{lemma-base-change-rig-smooth-homomorphism}
Let $A_1 \to A_2$ be a map of Noetherian rings. Let $I_i \subset A_i$
be an ideal such that $V(I_1A_2) = V(I_2)$. Let $B_1$ be in
(\ref{equation-C-prime}) for $(A_1, I_1)$.
Let $B_2$ be the base change of $B_1$ as in
Remark \ref{remark-base-change}.
If $B_1$ is rig-smooth over $(A_1, I_1)$,
then $B_2$ is rig-smooth over $(A_2, I_2)$.
\end{lemma}
\begin{proof}
Follows from Lemma \ref{lemma-zero-ext-1-after-modding-out} and
Definition \ref{definition-rig-smooth-homomorphism}
and the fact that $I_2^c$ is contained in $I_1A_2$ for some $c \geq 0$
as $A_2$ is Noetherian.
\end{proof}
\section{Deformations of ring homomorphisms}
\label{section-defos-ring-maps}
\noindent
Some work on lifting ring homomorphisms from rig-smooth algebras.
\begin{remark}[Linear approximation]
\label{remark-linear-approximation}
Let $A$ be a ring and $I \subset A$ be a finitely generated ideal.
Let $C$ be an $I$-adically complete $A$-algebra.
Let $\psi : A[x_1, \ldots, x_r]^\wedge \to C$ be a continuous
$A$-algebra map. Suppose given $\delta_i \in C$, $i = 1, \ldots, r$.
Then we can consider
$$
\psi' : A[x_1, \ldots, x_r]^\wedge \to C,\quad
x_i \longmapsto \psi(x_i) + \delta_i
$$
see Formal Spaces, Remark \ref{formal-spaces-remark-universal-property}.
Then we have
$$
\psi'(g) = \psi(g) + \sum \psi(\partial g/\partial x_i)\delta_i + \xi
$$
with error term $\xi \in (\delta_i\delta_j)$. This follows by
writing $g$ as a power series and working term by term. Convergence
is automatic as the coefficients of $g$ tend to zero.
Details omitted.
\end{remark}
\begin{remark}[Lifting maps]
\label{remark-improve-homomorphism}
Let $A$ be a Noetherian ring and $I \subset A$ be an ideal.
Let $B$ be an object of (\ref{equation-C-prime}).
Let $C$ be an $I$-adically complete $A$-algebra.
Let $\psi_n : B \to C/I^nC$ be an $A$-algebra homomorphism.
The obstruction to lifting $\psi_n$ to an $A$-algebra
homomorphism into $C/I^{2n}C$ is an element
$$
o(\psi_n) \in \Ext^1_B(\NL_{B/A}^\wedge, I^nC/I^{2n}C)
$$
as we will explain. Namely, choose a presentation
$B = A[x_1, \ldots, x_r]^\wedge/J$.
Choose a lift $\psi : A[x_1, \ldots, x_r]^\wedge \to C$ of $\psi_n$.
Since $\psi(J) \subset I^nC$ we get $\psi(J^2) \subset I^{2n}C$
and hence we get a $B$-linear homomorphism
$$
o(\psi) :
J/J^2 \longrightarrow I^nC/I^{2n}C, \quad g \longmapsto \psi(g)
$$
which of course extends to a $C$-linear map
$J/J^2 \otimes_B C \to I^nC/I^{2n}C$.
Since $\NL_{B/A}^\wedge = (J/J^2 \to \bigoplus B \text{d}x_i)$
we get $o(\psi_n)$ as the image of $o(\psi)$ by the identification
\begin{align*}
& \Ext^1_B(\NL_{B/A}^\wedge, I^nC/I^{2n}C) \\
& =
\Coker\left(\Hom_B(\bigoplus B\text{d}x_i, I^nC/I^{2n}C) \to
\Hom_B(J/J^2, I^nC/I^{2n}C)\right)
\end{align*}
See More on Algebra, Lemma
\ref{more-algebra-lemma-map-out-of-almost-free} part (1) for the equality.
\medskip\noindent
Suppose that $o(\psi_n)$ maps to zero in
$\Ext^1_B(\NL_{B/A}^\wedge, I^{n'}C/I^{2n'}C)$
for some integer $n'$ with $n > n' > n/2$. We claim that this means we can
find an $A$-algebra homomorphism $\psi'_{2n'} : B \to C/I^{2n'}C$
which agrees with $\psi_n$ as maps into $C/I^{n'}C$.
The extreme case $n' = n$ explains why we previously said
$o(\psi_n)$ is the obstruction to lifting $\psi_n$ to $C/I^{2n}C$.
Proof of the claim: the hypothesis that $o(\psi_n)$ maps to zero tells us
we can find a $B$-module map
$$
h : \bigoplus B\text{d}x_i \longrightarrow I^{n'}C/I^{2n'}C
$$
such that $o(\psi)$ and $h \circ \text{d}$ agree as maps
into $I^{n'}C/I^{2n'}C$. Say $h(\text{d}x_i) = \delta_i \bmod I^{2n'}C$
for some $\delta_i \in I^{n'}C$. Then we look at the map
$$
\psi' : A[x_1, \ldots, x_r]^\wedge \to C,\quad
x_i \longmapsto \psi(x_i) - \delta_i
$$
A computation with power series
shows that $\psi'(J) \subset I^{2n'}C$. Namely, for $g \in J$ we get
$$
\psi'(g) \equiv
\psi(g) - \sum \psi(\partial g/\partial x_i)\delta_i \equiv
o(\psi)(g) - (h \circ \text{d})(g) \equiv
0 \bmod I^{2n'}C
$$
See Remark \ref{remark-linear-approximation} for the first equality.
Hence $\psi'$ induces an $A$-algebra homomorphism
$\psi'_{2n'} : B \to C/I^{2n'}C$ as desired.
\end{remark}
\begin{lemma}
\label{lemma-get-morphism-general-better}
Assume given the following data
\begin{enumerate}
\item an integer $c \geq 0$,
\item an ideal $I$ of a Noetherian ring $A$,
\item $B$ in (\ref{equation-C-prime}) for $(A, I)$ such that
$I^c$ annihilates $\Ext^1_B(\NL_{B/A}^\wedge, N)$
for any $B$-module $N$,
\item a Noetherian $I$-adically complete $A$-algebra $C$; denote
$d = d(\text{Gr}_I(C))$ and $q_0 = q(\text{Gr}_I(C))$ the integers found in
Local Cohomology, Section \ref{local-cohomology-section-uniform},
\item an integer $n \geq \max(q_0 + (d + 1)c, 2(d + 1)c + 1)$, and
\item an $A$-algebra homomorphism $\psi_n : B \to C/I^nC$.
\end{enumerate}
Then there exists a map $\varphi : B \to C$ of $A$-algebras such
that $\psi_n \bmod I^{n - (d + 1)c} = \varphi \bmod I^{n - (d + 1)c}$.
\end{lemma}
\begin{proof}
Consider the obstruction class
$$
o(\psi_n) \in \Ext^1_B(\NL_{B/A}^\wedge, I^nC/I^{2n}C)
$$
of Remark \ref{remark-improve-homomorphism}. For any $C/I^nC$-module
$N$ we have
\begin{align*}
\Ext^1_B(\NL_{B/A}^\wedge, N)
& =
\Ext^1_{C/I^nC}(\NL_{B/A}^\wedge \otimes_B^\mathbf{L} C/I^nC, N) \\
& =
\Ext^1_{C/I^nC}(\NL_{B/A}^\wedge \otimes_B C/I^nC, N)
\end{align*}
The first equality by
More on Algebra, Lemma \ref{more-algebra-lemma-upgrade-adjoint-tensor-RHom}
and the second one by
More on Algebra, Lemma \ref{more-algebra-lemma-two-term-base-change}.
In particular, we see that
$\Ext^1_{C/I^nC}(\NL_{B/A}^\wedge \otimes_B C/I^nC, N)$ is annihilated by
$I^cC$ for all $C/I^nC$-modules $N$.
It follows that we may apply
Local Cohomology, Lemma \ref{local-cohomology-lemma-bound-two-term-complex}
to see that $o(\psi_n)$ maps to zero in
$$
\Ext^1_{C/I^nC}(\NL_{B/A}^\wedge \otimes_B C/I^nC, I^{n'}C/I^{2n'}C) =
\Ext^1_B(\NL_{B/A}^\wedge, I^{n'}C/I^{2n'}C) =
$$
where $n' = n - (d + 1)c$. By the discussion in