diff --git a/backend/audit/cross_validation/naming.py b/backend/audit/cross_validation/naming.py
index ed8885078b..de31a73966 100644
--- a/backend/audit/cross_validation/naming.py
+++ b/backend/audit/cross_validation/naming.py
@@ -1,21 +1,7 @@
+from types import new_class
 from typing import NamedTuple
 
-# There's no way around it, we need a canonical source of the different versions of each
-# name.
-"""
-    Starting point:
-
-        "AdditionalUEIs": "additional_ueis",
-        "AdditionalUEIs": "additional_ueis",
-        "AuditInformation": "audit_information",
-        "CorrectiveActionPlan": "corrective_action_plan",
-        "FederalAwards": "federal_awards",
-        "FindingsText": "findings_text",
-        "FindingsUniformGuidance": "findings_uniform_guidance",
-        "GeneralInformation": "general_information",
-        "NotesToSefa": "notes_to_sefa",
-        "TribalDataConsent": "tribal_data_consent",
-"""
+# We need a canonical source of the different versions of each name.
 
 
 class SectionBabelFish(NamedTuple):
@@ -157,6 +143,31 @@ class SectionBabelFish(NamedTuple):
 }
 
 
+# The following sets up NameConstantMetaclass as a metaclass that has property methods
+# for all of its attributes, then creates NC with NameConstantMetaclass as its
+# metaclass. The resulting class has each of the all_caps names from SECTION_NAMES as a
+# class-level attribute that is also read-only.
+# Thus e.g. NC.GENERAL_INFORMATION will return "general_information" and be read-only.
+
+
+def _readonly(value):
+    """
+    Given a value, return property method for that value, in order to set up the
+    metaclass below.
+    """
+    return property(lambda self: value)
+
+
+nameproperties = {
+    guide.all_caps: _readonly(guide.snake_case) for guide in SECTION_NAMES.values()
+}
+
+
+NameConstantsMetaclass = type("NCM", (type,), nameproperties)
+NC = new_class("NC", (object,), {"metaclass": NameConstantsMetaclass})
+
+
+# Helper funtions:
 def find_section_by_name(name):
     """
     Find the answers, first trying snake_case and then all the other versions.
diff --git a/backend/audit/cross_validation/sac_validation_shape.py b/backend/audit/cross_validation/sac_validation_shape.py
index 58ebe36804..0edbc19aff 100644
--- a/backend/audit/cross_validation/sac_validation_shape.py
+++ b/backend/audit/cross_validation/sac_validation_shape.py
@@ -1,10 +1,11 @@
 from audit.cross_validation.naming import (
+    NC,
     SECTION_NAMES,
     camel_to_snake,
     snake_to_camel,
 )
 
-at_root_sections = ("audit_information", "general_information")
+at_root_sections = (NC.AUDIT_INFORMATION, NC.GENERAL_INFORMATION)  # type: ignore
 
 
 def get_shaped_section(sac, section_name):
@@ -34,7 +35,7 @@ def sac_validation_shape(sac):
     that's appropriate for passing to the validation functions.
 
     For example, if the Audit Information and Notes to SEFA sections have content,
-    this function wil return something like:
+    this function will return something like:
 
     {
         "sf_sac_sections": {
diff --git a/backend/audit/cross_validation/submission_progress_check.py b/backend/audit/cross_validation/submission_progress_check.py
index 2f7a161687..07165d6f65 100644
--- a/backend/audit/cross_validation/submission_progress_check.py
+++ b/backend/audit/cross_validation/submission_progress_check.py
@@ -1,4 +1,4 @@
-from audit.cross_validation.naming import find_section_by_name
+from audit.cross_validation.naming import NC, find_section_by_name
 
 
 def submission_progress_check(sac, sar=None, crossval=True):
@@ -36,10 +36,10 @@ def submission_progress_check(sac, sar=None, crossval=True):
         }
     """
     # TODO: remove these once tribal data consent are implemented:
-    del sac["sf_sac_sections"]["tribal_data_consent"]
+    del sac["sf_sac_sections"][NC.TRIBAL_DATA_CONSENT]
 
     # Add the status of the SAR into the list of sections:
-    sac["sf_sac_sections"]["single_audit_report"] = bool(sar)
+    sac["sf_sac_sections"][NC.SINGLE_AUDIT_REPORT] = bool(sar)
 
     result = {k: None for k in sac["sf_sac_sections"]}
 
@@ -87,22 +87,22 @@ def get_num_findings(award):
         "section_name": key,
     }
     awards = {}
-    if sections["federal_awards"]:
-        awards = sections.get("federal_awards", {}).get("federal_awards", [])
-    general_info = sections.get("general_information", {}) or {}
+    if sections[NC.FEDERAL_AWARDS]:
+        awards = sections.get(NC.FEDERAL_AWARDS, {}).get(NC.FEDERAL_AWARDS, [])
+    general_info = sections.get(NC.GENERAL_INFORMATION, {}) or {}
     num_findings = sum(get_num_findings(award) for award in awards)
     conditions = {
-        "general_information": True,
-        "audit_information": True,
-        "federal_awards": True,
-        "notes_to_sefa": True,
-        "findings_uniform_guidance": num_findings > 0,
-        "findings_text": num_findings > 0,
-        "corrective_action_plan": num_findings > 0,
-        "additional_ueis": bool(general_info.get("multiple_ueis_covered")),
-        "additional_eins": bool(general_info.get("multiple_eins_covered")),
-        "secondary_auditors": bool(general_info.get("secondary_auditors_exist")),
-        "single_audit_report": True,
+        NC.GENERAL_INFORMATION: True,
+        NC.AUDIT_INFORMATION: True,
+        NC.FEDERAL_AWARDS: True,
+        NC.NOTES_TO_SEFA: True,
+        NC.FINDINGS_UNIFORM_GUIDANCE: num_findings > 0,
+        NC.FINDINGS_TEXT: num_findings > 0,
+        NC.CORRECTIVE_ACTION_PLAN: num_findings > 0,
+        NC.ADDITIONAL_UEIS: bool(general_info.get("multiple_ueis_covered")),
+        NC.ADDITIONAL_EINS: bool(general_info.get("multiple_eins_covered")),
+        NC.SECONDARY_AUDITORS: bool(general_info.get("secondary_auditors_exist")),
+        NC.SINGLE_AUDIT_REPORT: True,
     }
 
     # If it's not required, it's inactive: