diff --git a/ci/conda_env_r.txt b/ci/conda_env_r.txt
index 03d5f3b625c8..0ac6b1879db3 100644
--- a/ci/conda_env_r.txt
+++ b/ci/conda_env_r.txt
@@ -15,7 +15,6 @@
 # specific language governing permissions and limitations
 # under the License.
 
-r-assertthat
 r-base
 r-bit64
 r-dplyr
diff --git a/r/DESCRIPTION b/r/DESCRIPTION
index f682de189e72..54a1d00f894a 100644
--- a/r/DESCRIPTION
+++ b/r/DESCRIPTION
@@ -33,13 +33,12 @@ SystemRequirements: C++20; for AWS S3 support on Linux, libcurl and openssl, and
     cmake >= 3.26 (build-time only, and only for full source build)
 Biarch: true
 Imports:
-    assertthat,
     bit64 (>= 0.9-7),
     glue,
     methods,
     purrr,
     R6,
-    rlang (>= 1.0.0),
+    rlang (>= 1.2.0),
     stats,
     tidyselect (>= 1.0.0),
     utils,
diff --git a/r/NAMESPACE b/r/NAMESPACE
index 46c29b3e9370..b1648ba7f50b 100644
--- a/r/NAMESPACE
+++ b/r/NAMESPACE
@@ -428,8 +428,6 @@ export(write_parquet)
 export(write_to_raw)
 export(write_tsv_dataset)
 importFrom(R6,R6Class)
-importFrom(assertthat,assert_that)
-importFrom(assertthat,is.string)
 importFrom(bit64,integer64)
 importFrom(glue,glue)
 importFrom(methods,as)
@@ -460,9 +458,14 @@ importFrom(rlang,as_quosure)
 importFrom(rlang,call2)
 importFrom(rlang,call_args)
 importFrom(rlang,call_name)
+importFrom(rlang,caller_arg)
 importFrom(rlang,caller_env)
+importFrom(rlang,check_bool)
 importFrom(rlang,check_dots_empty)
 importFrom(rlang,check_dots_empty0)
+importFrom(rlang,check_number_decimal)
+importFrom(rlang,check_number_whole)
+importFrom(rlang,check_string)
 importFrom(rlang,dots_list)
 importFrom(rlang,dots_n)
 importFrom(rlang,enexpr)
@@ -482,15 +485,22 @@ importFrom(rlang,is_bare_character)
 importFrom(rlang,is_bare_list)
 importFrom(rlang,is_call)
 importFrom(rlang,is_character)
+importFrom(rlang,is_closure)
 importFrom(rlang,is_empty)
+importFrom(rlang,is_environment)
 importFrom(rlang,is_false)
 importFrom(rlang,is_formula)
 importFrom(rlang,is_integerish)
 importFrom(rlang,is_interactive)
 importFrom(rlang,is_list)
+importFrom(rlang,is_logical)
+importFrom(rlang,is_missing)
+importFrom(rlang,is_na)
+importFrom(rlang,is_null)
 importFrom(rlang,is_quosure)
 importFrom(rlang,is_string)
 importFrom(rlang,is_symbol)
+importFrom(rlang,is_vector)
 importFrom(rlang,list2)
 importFrom(rlang,new_data_mask)
 importFrom(rlang,new_environment)
@@ -508,6 +518,7 @@ importFrom(rlang,quo_set_expr)
 importFrom(rlang,quos)
 importFrom(rlang,seq2)
 importFrom(rlang,set_names)
+importFrom(rlang,stop_input_type)
 importFrom(rlang,sym)
 importFrom(rlang,syms)
 importFrom(rlang,trace_back)
diff --git a/r/NEWS.md b/r/NEWS.md
index c8588cb9b4ae..10078dd8e417 100644
--- a/r/NEWS.md
+++ b/r/NEWS.md
@@ -19,6 +19,9 @@
 
 # arrow 25.0.0.9000
 
+## Minor improvements and fixes
+- arrow no longer depends on assertthat and has impr (@olivroy, #50525).
+
 # arrow 25.0.0
 
 ## Breaking changes
diff --git a/r/R/array.R b/r/R/array.R
index f2b34fc03f8b..d9a4c9d93a9f 100644
--- a/r/R/array.R
+++ b/r/R/array.R
@@ -417,7 +417,9 @@ DictionaryArray$create <- function(x, dict = NULL) {
     return(Array$create(x))
   }
 
-  assert_that(!is.null(dict))
+  if (is.null(dict)) {
+    stop("dict cannot be `NULL`.", call. = FALSE)
+  }
   if (!is.Array(x)) {
     x <- Array$create(x)
   }
@@ -450,18 +452,17 @@ StructArray$create <- function(...) {
 
 #' @export
 `[[.StructArray` <- function(x, i, ...) {
+  assert_is(i, c("character", "numeric", "integer"))
   if (is.character(i)) {
     x$GetFieldByName(i)
   } else if (is.numeric(i)) {
     x$field(i - 1)
-  } else {
-    stop("'i' must be character or numeric, not ", class(i), call. = FALSE)
   }
 }
 
 #' @export
 `$.StructArray` <- function(x, name, ...) {
-  assert_that(is.string(name))
+  check_string(name)
   if (name %in% ls(x)) {
     get(name, x)
   } else {
diff --git a/r/R/arrow-datum.R b/r/R/arrow-datum.R
index 5ac72bf36856..fa6918b7ed83 100644
--- a/r/R/arrow-datum.R
+++ b/r/R/arrow-datum.R
@@ -29,9 +29,7 @@ ArrowDatum <- R6Class(
       call_function("cast", self, options = opts)
     },
     SortIndices = function(descending = FALSE) {
-      assert_that(is.logical(descending))
-      assert_that(length(descending) == 1L)
-      assert_that(!is.na(descending))
+      check_bool(descending, allow_na = FALSE)
       call_function(
         "sort_indices",
         self,
@@ -285,8 +283,7 @@ filter_rows <- function(x, i, keep_na = TRUE, ...) {
 #' @importFrom utils head
 #' @export
 head.ArrowDatum <- function(x, n = 6L, ...) {
-  assert_is(n, c("numeric", "integer"))
-  assert_that(length(n) == 1)
+  check_number_decimal(n)
   len <- NROW(x)
   if (n < 0) {
     # head(x, negative) means all but the last n rows
@@ -306,8 +303,7 @@ head.ArrowDatum <- function(x, n = 6L, ...) {
 #' @importFrom utils tail
 #' @export
 tail.ArrowDatum <- function(x, n = 6L, ...) {
-  assert_is(n, c("numeric", "integer"))
-  assert_that(length(n) == 1)
+  check_number_decimal(n)
   if (!is.integer(n)) {
     n <- floor(n)
   }
diff --git a/r/R/arrow-package.R b/r/R/arrow-package.R
index 2706faee5cb1..3ca90bde0e0b 100644
--- a/r/R/arrow-package.R
+++ b/r/R/arrow-package.R
@@ -19,7 +19,6 @@
 #' @importFrom R6 R6Class
 #' @importFrom purrr as_mapper map map2 map_chr map2_chr map_dbl map_dfr map_int map_lgl keep imap imap_chr
 #' @importFrom purrr compact flatten reduce walk
-#' @importFrom assertthat assert_that is.string
 #' @importFrom rlang list2 %||% is_false abort dots_n warn enquo quo_is_null enquos is_integerish quos quo
 #' @importFrom rlang eval_tidy new_data_mask syms env new_environment env_bind set_names exec
 #' @importFrom rlang is_bare_character quo_get_expr quo_get_env quo_set_expr .data seq2 is_interactive
@@ -27,7 +26,9 @@
 #' @importFrom rlang is_list call2 is_empty as_function as_label arg_match is_symbol is_call call_args
 #' @importFrom rlang quo_set_env quo_get_env is_formula quo_is_call f_rhs parse_expr f_env new_quosure
 #' @importFrom rlang new_quosures expr_text caller_env check_dots_empty check_dots_empty0 dots_list is_string inform
-#' @importFrom rlang is_bare_list call_name
+#' @importFrom rlang is_bare_list call_name check_string check_number_whole check_number_decimal check_bool
+#' @importFrom rlang is_logical caller_env caller_arg stop_input_type is_null is_closure
+#' @importFrom rlang is_environment is_na is_vector is_missing
 #' @importFrom tidyselect vars_pull eval_select eval_rename
 #' @importFrom glue glue
 #' @importFrom bit64 integer64
diff --git a/r/R/arrow-tabular.R b/r/R/arrow-tabular.R
index a83226a728c3..653da6f8ac54 100644
--- a/r/R/arrow-tabular.R
+++ b/r/R/arrow-tabular.R
@@ -36,26 +36,34 @@ ArrowTabular <- R6Class(
       if (is.integer(i)) {
         i <- Array$create(i)
       }
-      assert_that(is.Array(i))
+      if (!is.Array(i)) {
+        stop("`i` must be an Array.")
+      }
       call_function("take", self, i)
     },
     Filter = function(i, keep_na = TRUE) {
       if (is.logical(i)) {
         i <- Array$create(i)
       }
-      assert_that(is.Array(i, "bool"))
+      if (!is.Array(i, "bool")) {
+        stop("i must be an Array of bool", call. = FALSE)
+      }
       call_function("filter", self, i, options = list(keep_na = keep_na))
     },
     SortIndices = function(names, descending = FALSE) {
-      assert_that(is.character(names))
-      assert_that(length(names) > 0)
-      assert_that(!anyNA(names))
+      # if rlang adds allow_empty argument https://github.com/r-lib/rlang/pull/1745
+      # use allow_empty = FALSE instead
+      check_character(names, allow_na = FALSE)
+      if (length(names) == 0) {
+        stop("names can't be empty.", call. = FALSE)
+      }
       if (length(descending) == 1L) {
         descending <- rep_len(descending, length(names))
       }
-      assert_that(is.logical(descending))
-      assert_that(identical(length(names), length(descending)))
-      assert_that(!anyNA(descending))
+      check_logical(descending, allow_na = FALSE)
+      if (length(names) != length(descending)) {
+        stop("names and descending must have a common length", call. = FALSE)
+      }
       call_function(
         "sort_indices",
         self,
@@ -142,18 +150,17 @@ as.data.frame.ArrowTabular <- function(x, row.names = NULL, optional = FALSE, ..
 
 #' @export
 `[[.ArrowTabular` <- function(x, i, ...) {
+  assert_is(i, c("character", "numeric"))
   if (is.character(i)) {
     x$GetColumnByName(i)
   } else if (is.numeric(i)) {
     x$column(i - 1)
-  } else {
-    stop("'i' must be character or numeric, not ", class(i), call. = FALSE)
   }
 }
 
 #' @export
 `$.ArrowTabular` <- function(x, name, ...) {
-  assert_that(is.string(name))
+  check_string(name)
   if (name %in% ls(x)) {
     get(name, x)
   } else {
@@ -163,10 +170,10 @@ as.data.frame.ArrowTabular <- function(x, row.names = NULL, optional = FALSE, ..
 
 #' @export
 `[[<-.ArrowTabular` <- function(x, i, value) {
-  if (!is.character(i) && !is.numeric(i)) {
-    stop("'i' must be character or numeric, not ", class(i), call. = FALSE)
+  assert_is(i, c("character", "numeric"))
+  if (length(i) != 1 || is.na(i)) {
+    stop("`i` must be length 1 and not NA", call. = FALSE)
   }
-  assert_that(length(i) == 1, !is.na(i))
 
   if (is.null(value)) {
     if (is.character(i)) {
@@ -209,7 +216,7 @@ as.data.frame.ArrowTabular <- function(x, row.names = NULL, optional = FALSE, ..
 
 #' @export
 `$<-.ArrowTabular` <- function(x, i, value) {
-  assert_that(is.string(i))
+  check_string(i)
   # We need to check if `i` is in names in case it is an active binding (e.g.
   # `metadata`, in which case we use assign to change the active binding instead
   # of the column in the table)
diff --git a/r/R/compression.R b/r/R/compression.R
index 4cd50d47a06f..3477f9cbd849 100644
--- a/r/R/compression.R
+++ b/r/R/compression.R
@@ -47,7 +47,7 @@ Codec <- R6Class(
   )
 )
 Codec$create <- function(type = "gzip", compression_level = NA) {
-  if (is.string(type)) {
+  if (is_string(type)) {
     type <- util___Codec__Create(
       compression_from_name(type),
       compression_level
@@ -104,7 +104,7 @@ compression_from_name <- function(name) {
 CompressedOutputStream <- R6Class("CompressedOutputStream", inherit = OutputStream)
 CompressedOutputStream$create <- function(stream, codec = "gzip", compression_level = NA) {
   codec <- Codec$create(codec, compression_level = compression_level)
-  if (is.string(stream)) {
+  if (is_string(stream)) {
     stream <- FileOutputStream$create(stream)
   }
   assert_is(stream, "OutputStream")
@@ -118,7 +118,7 @@ CompressedOutputStream$create <- function(stream, codec = "gzip", compression_le
 CompressedInputStream <- R6Class("CompressedInputStream", inherit = InputStream)
 CompressedInputStream$create <- function(stream, codec = "gzip", compression_level = NA) {
   codec <- Codec$create(codec, compression_level = compression_level)
-  if (is.string(stream)) {
+  if (is_string(stream)) {
     stream <- ReadableFile$create(stream)
   }
   assert_is(stream, "InputStream")
diff --git a/r/R/compute.R b/r/R/compute.R
index d5da9024bf46..b21d820127ac 100644
--- a/r/R/compute.R
+++ b/r/R/compute.R
@@ -44,8 +44,10 @@
 #' @include chunked-array.R
 #' @include scalar.R
 call_function <- function(function_name, ..., args = list(...), options = empty_named_list()) {
-  assert_that(is.string(function_name))
-  assert_that(is.list(options), !is.null(names(options)))
+  check_string(function_name)
+  if (!is.list(options) || is.null(names(options))) {
+    stop("options must be a named list.", call. = FALSE)
+  }
 
   datum_classes <- c("Array", "ChunkedArray", "RecordBatch", "Table", "Scalar")
   valid_args <- map_lgl(args, ~ inherits(., datum_classes))
@@ -152,7 +154,9 @@ collect_arrays_from_dots <- function(dots) {
     return(dots[[1]])
   }
 
-  assert_that(all(map_lgl(dots, is.Array)))
+  if (!all(map_lgl(dots, is.Array))) {
+    stop("dots must all be Arrays.", call. = FALSE)
+  }
   arrays <- unlist(lapply(dots, function(x) {
     if (inherits(x, "ChunkedArray")) {
       x$chunks
@@ -176,8 +180,10 @@ quantile.ArrowDatum <- function(
     x <- Array$create(x)
   }
   assert_is(probs, c("numeric", "integer"))
-  assert_that(length(probs) > 0)
-  assert_that(all(probs >= 0 & probs <= 1))
+  # if check_numbers_decimal() gets added in rlang, https://github.com/r-lib/rlang/issues/1714
+  if (length(probs) == 0 || !all(probs >= 0 & probs <= 1)) {
+    stop("probs must be a probability vector between 0 and 1.")
+  }
   if (!na.rm && x$null_count > 0) {
     stop("Missing values not allowed if 'na.rm' is FALSE", call. = FALSE)
   }
diff --git a/r/R/csv.R b/r/R/csv.R
index a0ecd48677a8..29362219797d 100644
--- a/r/R/csv.R
+++ b/r/R/csv.R
@@ -440,7 +440,7 @@ csv_read_options <- function(
   encoding = "UTF-8",
   skip_rows_after_names = 0L
 ) {
-  assert_that(is.string(encoding))
+  check_string(encoding)
 
   options <- csv___ReadOptions__initialize(
     list(
@@ -646,14 +646,14 @@ csv_write_options <- function(
   quoting_style_opts <- c("Needed", "AllValid", "None")
   quoting_style <- match(quoting_style, quoting_style_opts) - 1L
 
-  assert_that(is.logical(include_header))
-  assert_that(is_integerish(batch_size, n = 1, finite = TRUE), batch_size > 0)
-  assert_that(is.character(delimiter))
-  assert_that(is.character(null_string))
-  assert_that(!is.na(null_string))
-  assert_that(length(null_string) == 1)
-  assert_that(!grepl('"', null_string), msg = "na argument must not contain quote characters.")
-  assert_that(is.character(eol))
+  check_bool(include_header)
+  check_number_whole(batch_size, min = 1, allow_infinite = FALSE)
+  check_string(delimiter)
+  check_string(null_string, allow_na = FALSE)
+  if (grepl('"', null_string)) {
+    stop("na argument must not contain quote characters.", call. = FALSE)
+  }
+  check_string(eol)
 
   csv___WriteOptions__initialize(
     list(
@@ -930,17 +930,17 @@ write_csv_arrow <- function(
   write_options = NULL,
   ...
 ) {
-  unsupported_passed_args <- names(list(...))
-
-  if (length(unsupported_passed_args)) {
-    stop(
-      "The following ",
-      ngettext(length(unsupported_passed_args), "argument is ", "arguments are "),
-      "not yet supported in Arrow: ",
-      oxford_paste(unsupported_passed_args),
-      call. = FALSE
+  # every other argument is not supported in arrow and inform those arguments
+  # are not yet supported in arrow.
+  check_dots_empty(error = function(cnd) {
+    rlang::abort(
+      c(
+        "Arguments not yet supported in Arrow",
+        conditionMessage(cnd)
+      ),
+      call = call("write_csv_arrow")
     )
-  }
+  })
 
   if (!missing(file) && !missing(sink)) {
     stop(
diff --git a/r/R/dataset-factory.R b/r/R/dataset-factory.R
index 02b2b8553c19..9e54c17ed94a 100644
--- a/r/R/dataset-factory.R
+++ b/r/R/dataset-factory.R
@@ -233,18 +233,14 @@ FileSystemDatasetFactory$create <- function(
   factory_options = list()
 ) {
   assert_is(filesystem, "FileSystem")
-  is.null(selector) || assert_is(selector, "FileSelector")
-  is.null(paths) || assert_is(paths, "character")
-  assert_that(
-    xor(is.null(selector), is.null(paths)),
-    msg = "Either selector or paths must be specified"
-  )
+  assert_is(selector, "FileSelector", allow_null = TRUE)
+  assert_is(paths, "character", allow_null = TRUE)
+  if (!xor(is.null(selector), is.null(paths))) {
+    stop("Either selector or paths must be specified")
+  }
   assert_is(format, "FileFormat")
   if (!is.null(paths)) {
-    assert_that(
-      is.null(partitioning),
-      msg = "Partitioning not supported with paths"
-    )
+    check_null(partitioning, msg = "Partitioning not supported with paths")
     # Validate that exclude_invalid_files is only option provided
     # All other options are only relevant for the FileSelector method
     invalid_opts <- setdiff(names(factory_options), "exclude_invalid_files")
@@ -285,15 +281,14 @@ fsf_options <- function(factory_options, partitioning) {
   if (!is.null(factory_options$partition_base_dir)) {
     if (
       inherits(partitioning, "HivePartitioning") ||
-        (inherits(partitioning, "PartitioningFactory") &&
-          identical(partitioning$type_name, "hive"))
+        (inherits(partitioning, "PartitioningFactory") && identical(partitioning$type_name, "hive"))
     ) {
       warning(
         "factory_options$partition_base_dir is not meaningful for Hive partitioning",
         call. = FALSE
       )
     } else {
-      assert_that(is.string(factory_options$partition_base_dir))
+      check_string(factory_options$partition_base_dir)
     }
   }
 
diff --git a/r/R/dataset-scan.R b/r/R/dataset-scan.R
index c4b651d419dd..d56aa5c64e6f 100644
--- a/r/R/dataset-scan.R
+++ b/r/R/dataset-scan.R
@@ -160,10 +160,8 @@ names.Scanner <- function(x) names(x$schema)
 
 #' @export
 head.Scanner <- function(x, n = 6L, ...) {
-  assert_is(n, c("numeric", "integer"))
-  assert_that(length(n) == 1)
   # Negative n requires knowing nrow(x), which requires a scan itself
-  assert_that(n >= 0)
+  check_number_decimal(n, min = 0)
   if (!is.integer(n)) {
     n <- floor(n)
   }
@@ -176,10 +174,8 @@ tail.Scanner <- function(x, n = 6L, ...) {
 }
 
 tail_from_batches <- function(batches, n) {
-  assert_is(n, c("numeric", "integer"))
-  assert_that(length(n) == 1)
   # Negative n requires knowing nrow(x), which requires a scan itself
-  assert_that(n >= 0)
+  check_number_decimal(n, min = 0)
   if (!is.integer(n)) {
     n <- floor(n)
   }
diff --git a/r/R/dplyr-eval.R b/r/R/dplyr-eval.R
index 1282f171878d..f85b63fc2370 100644
--- a/r/R/dplyr-eval.R
+++ b/r/R/dplyr-eval.R
@@ -47,7 +47,7 @@ arrow_eval <- function(expr, mask) {
       stop(e)
     }
 
-    # 2. Error is from assert_that: raise as validation_error
+    # 2. Error is from assert_is: raise as validation_error
     if (inherits(e, "assertError")) {
       validation_error(msg, call = expr)
     }
@@ -171,7 +171,7 @@ i18ize_error_messages <- function() {
 #' suggestion, when the error is ultimately raised by `try_error_dplyr()`,
 #' `Call collect() first to pull data into R` won't be the only suggestion.
 #'
-#' You can still use `match.arg()` and `assert_that()` for simple input
+#' You can still use `match.arg()` and `assert_is()` for simple input
 #' validation inside of the function bindings. `arrow_eval()` will catch their
 #' errors and re-raise them as `validation_error`.
 #'
diff --git a/r/R/dplyr-funcs-string.R b/r/R/dplyr-funcs-string.R
index 158bae2db87c..1cfe350dd2cf 100644
--- a/r/R/dplyr-funcs-string.R
+++ b/r/R/dplyr-funcs-string.R
@@ -292,7 +292,7 @@ register_bindings_string_regex <- function() {
     "stringr::str_count",
     function(string, pattern) {
       opts <- get_stringr_pattern_options(enquo(pattern))
-      if (!is.string(pattern)) {
+      if (!is_string(pattern)) {
         arrow_not_supported("`pattern` must be a length 1 character vector; other values")
       }
       arrow_fun <- ifelse(opts$fixed, "count_substring", "count_substring_regex")
@@ -422,7 +422,7 @@ register_bindings_string_regex <- function() {
   })
 
   register_binding("base::strsplit", function(x, split, fixed = FALSE, perl = FALSE, useBytes = FALSE) {
-    assert_that(is.string(split))
+    check_string(split)
 
     arrow_fun <- ifelse(fixed, "split_pattern", "split_pattern_regex")
     # warn when the user specifies both fixed = TRUE and perl = TRUE, for
@@ -620,9 +620,11 @@ register_bindings_string_other <- function() {
   )
 
   register_binding("stringr::str_pad", function(string, width, side = c("left", "right", "both"), pad = " ") {
-    assert_that(is_integerish(width))
+    if (!is_integerish(width)) {
+      abort("width must be whole numbers.")
+    }
     side <- match.arg(side)
-    assert_that(is.string(pad))
+    check_string(pad)
 
     if (side == "left") {
       pad_func <- "utf8_lpad"
diff --git a/r/R/dplyr-funcs-type.R b/r/R/dplyr-funcs-type.R
index 0c8864ac6c77..e589900bc766 100644
--- a/r/R/dplyr-funcs-type.R
+++ b/r/R/dplyr-funcs-type.R
@@ -78,7 +78,7 @@ register_bindings_type_cast <- function() {
   })
 
   register_binding("methods::is", function(object, class2) {
-    if (is.string(class2)) {
+    if (is_string(class2)) {
       switch(
         class2,
         # for R data types, pass off to is.*() functions
@@ -245,26 +245,26 @@ register_bindings_type_inspect <- function() {
 
   # rlang::is_* type functions
   register_binding("rlang::is_character", function(x, n = NULL) {
-    assert_that(is.null(n))
+    check_null(n)
     call_binding("is.character", x)
   })
   register_binding("rlang::is_double", function(x, n = NULL, finite = NULL) {
-    assert_that(is.null(n))
+    check_null(n)
     if (!is.null(finite)) {
       arrow_not_supported("`finite` argument")
     }
     call_binding("is.double", x)
   })
   register_binding("rlang::is_integer", function(x, n = NULL) {
-    assert_that(is.null(n))
+    check_null(n)
     call_binding("is.integer", x)
   })
   register_binding("rlang::is_list", function(x, n = NULL) {
-    assert_that(is.null(n))
+    check_null(n)
     call_binding("is.list", x)
   })
   register_binding("rlang::is_logical", function(x, n = NULL) {
-    assert_that(is.null(n))
+    check_null(n)
     call_binding("is.logical", x)
   })
 }
@@ -277,8 +277,7 @@ register_bindings_type_elementwise <- function() {
   register_binding("base::is.nan", function(x) {
     if (
       is.double(x) ||
-        (inherits(x, "Expression") &&
-          x$type_id() %in% TYPES_WITH_NAN)
+        (inherits(x, "Expression") && x$type_id() %in% TYPES_WITH_NAN)
     ) {
       # TODO: if an option is added to the is_nan kernel to treat NA as NaN,
       # use that to simplify the code here (ARROW-13366)
@@ -312,11 +311,8 @@ register_bindings_type_format <- function() {
     if (!inherits(x, "Expression")) {
       return(format(x, ...))
     }
-
-    if (
-      inherits(x, "Expression") &&
-        x$type_id() %in% Type[c("TIMESTAMP", "DATE32", "DATE64")]
-    ) {
+    accepted_types <- Type[c("TIMESTAMP", "DATE32", "DATE64")]
+    if (inherits(x, "Expression") && x$type_id() %in% accepted_types) {
       binding_format_datetime(x, ...)
     } else {
       cast(x, string())
diff --git a/r/R/dplyr-summarize.R b/r/R/dplyr-summarize.R
index 7c2a44eec3d5..2e63373ef5d8 100644
--- a/r/R/dplyr-summarize.R
+++ b/r/R/dplyr-summarize.R
@@ -146,7 +146,7 @@ do_arrow_summarize <- function(.data, ..., .groups = NULL) {
       # But we don't support anything that returns multiple rows now
       .groups <- "drop_last"
     } else {
-      assert_that(is.string(.groups))
+      check_string(.groups)
     }
     if (.groups == "drop_last") {
       out$group_by_vars <- head(.data$group_by_vars, -1)
diff --git a/r/R/dplyr.R b/r/R/dplyr.R
index 097a48489abc..2c68c7cca9d8 100644
--- a/r/R/dplyr.R
+++ b/r/R/dplyr.R
@@ -230,8 +230,7 @@ as.data.frame.arrow_dplyr_query <- function(x, row.names = NULL, optional = FALS
 
 #' @export
 head.arrow_dplyr_query <- function(x, n = 6L, ...) {
-  assert_is(n, c("numeric", "integer"))
-  assert_that(length(n) == 1)
+  check_number_decimal(n)
   if (!is.integer(n)) {
     n <- floor(n)
   }
@@ -241,8 +240,7 @@ head.arrow_dplyr_query <- function(x, n = 6L, ...) {
 
 #' @export
 tail.arrow_dplyr_query <- function(x, n = 6L, ...) {
-  assert_is(n, c("numeric", "integer"))
-  assert_that(length(n) == 1)
+  check_number_decimal(n)
   if (!is.integer(n)) {
     n <- floor(n)
   }
diff --git a/r/R/expression.R b/r/R/expression.R
index 78272323c709..3981b43a8971 100644
--- a/r/R/expression.R
+++ b/r/R/expression.R
@@ -51,11 +51,15 @@ Expression <- R6Class(
     # schemas in Expression objects (ARROW-13186)
     schema = NULL,
     type = function(schema = self$schema) {
-      assert_that(!is.null(schema))
+      if (is.null(schema)) {
+        stop("schema cannot be NULL.", call. = FALSE)
+      }
       compute___expr__type(self, schema)
     },
     type_id = function(schema = self$schema) {
-      assert_that(!is.null(schema))
+      if (is.null(schema)) {
+        stop("schema cannot be NULL", call. = FALSE)
+      }
       compute___expr__type_id(self, schema)
     },
     is_field_ref = function() {
@@ -75,7 +79,7 @@ Expression <- R6Class(
   )
 )
 Expression$create <- function(function_name, ..., args = list(...), options = empty_named_list()) {
-  assert_that(is.string(function_name))
+  check_string(function_name)
   # Make sure all inputs are Expressions
   args <- lapply(args, function(x) {
     if (!inherits(x, "Expression")) {
@@ -99,7 +103,7 @@ Expression$create <- function(function_name, ..., args = list(...), options = em
 
 #' @export
 `$.Expression` <- function(x, name, ...) {
-  assert_that(is.string(name))
+  check_string(name)
   if (name %in% ls(x)) {
     get(name, x)
   } else {
@@ -111,7 +115,7 @@ get_nested_field <- function(expr, name) {
   if (expr$is_field_ref()) {
     # Make a nested field ref
     # TODO(#33756): integer (positional) field refs are supported in C++
-    assert_that(is.string(name))
+    check_string(name)
     out <- compute___expr__nested_field_ref(expr, name)
   } else {
     # Use the struct_field kernel if expr is a struct:
@@ -149,7 +153,7 @@ get_nested_field <- function(expr, name) {
 
 Expression$field_ref <- function(name) {
   # TODO(#33756): allow construction of field ref from integer
-  assert_that(is.string(name))
+  check_string(name)
   compute___expr__field_ref(name)
 }
 Expression$scalar <- function(x) {
diff --git a/r/R/extension.R b/r/R/extension.R
index 1fe073d7401f..cc9eb518702a 100644
--- a/r/R/extension.R
+++ b/r/R/extension.R
@@ -227,11 +227,13 @@ ExtensionType$new <- function(xp) {
 }
 
 ExtensionType$create <- function(storage_type, extension_name, extension_metadata = raw(), type_class = ExtensionType) {
-  if (is.string(extension_metadata)) {
+  if (is_string(extension_metadata)) {
     extension_metadata <- charToRaw(enc2utf8(extension_metadata))
   }
-
-  assert_that(is.string(extension_name), is.raw(extension_metadata))
+  check_string(extension_name)
+  if (!is.raw(extension_metadata)) {
+    stop("`extension_metadata` must be raw.", call. = FALSE)
+  }
   assert_is(storage_type, "DataType")
   assert_is(type_class, "R6ClassGenerator")
 
diff --git a/r/R/feather.R b/r/R/feather.R
index 23d5e4ed20c0..4c33c138fb12 100644
--- a/r/R/feather.R
+++ b/r/R/feather.R
@@ -76,7 +76,7 @@ write_feather <- function(
 ) {
   # Handle and validate options before touching data
   version <- as.integer(version)
-  assert_that(version %in% 1:2)
+  check_number_whole(version, min = 1, max = 2)
 
   if (isTRUE(compression)) {
     compression <- "default"
@@ -88,7 +88,7 @@ write_feather <- function(
   # TODO(ARROW-17221): if (missing(compression)), we could detect_compression(sink) here
   compression <- match.arg(compression)
   chunk_size <- as.integer(chunk_size)
-  assert_that(chunk_size > 0)
+  check_number_whole(chunk_size, min = 1)
   if (compression == "default") {
     if (version == 2 && codec_is_available("lz4")) {
       compression <- "lz4"
diff --git a/r/R/field.R b/r/R/field.R
index f1a9c8a2eac1..ec0e7d6c4a25 100644
--- a/r/R/field.R
+++ b/r/R/field.R
@@ -84,7 +84,7 @@ Field <- R6Class(
   )
 )
 Field$create <- function(name, type, metadata = NULL, nullable = TRUE) {
-  assert_that(inherits(name, "character"), length(name) == 1L)
+  check_string(name)
   type <- as_type(type, name)
   f <- Field__initialize(enc2utf8(name), type, nullable)
   if (!is.null(metadata)) {
@@ -113,7 +113,10 @@ field <- Field$create
 
 .fields <- function(.list, nullable = TRUE) {
   if (length(.list)) {
-    assert_that(!is.null(nms <- names(.list)))
+    nms <- names(.list)
+    if (is.null(nms)) {
+      stop(".list must be named.", call. = FALSE)
+    }
     map2(nms, .list, field)
   } else {
     list()
diff --git a/r/R/filesystem.R b/r/R/filesystem.R
index abbc9cd6c5bf..f7c80b092a77 100644
--- a/r/R/filesystem.R
+++ b/r/R/filesystem.R
@@ -367,7 +367,7 @@ FileSystem <- R6Class(
   )
 )
 FileSystem$from_uri <- function(uri) {
-  assert_that(is.string(uri))
+  check_string(uri)
   fs___FileSystemFromUri(uri)
 }
 
@@ -377,7 +377,9 @@ get_paths_and_filesystem <- function(x, filesystem = NULL) {
   if (inherits(x, "SubTreeFileSystem")) {
     return(list(fs = x$base_fs, path = x$base_path))
   }
-  assert_that(is.character(x))
+  if (!is.character(x)) {
+    stop("x must be a character vector.", call. = FALSE)
+  }
   are_urls <- are_urls(x)
   if (any(are_urls)) {
     if (!all(are_urls)) {
@@ -412,12 +414,14 @@ get_paths_and_filesystem <- function(x, filesystem = NULL) {
 # variant of the above function that asserts that x is either a scalar string
 # or a SubTreeFileSystem
 get_path_and_filesystem <- function(x, filesystem = NULL) {
-  assert_that(is.string(x) || inherits(x, "SubTreeFileSystem"))
+  if (!is_string(x) && !inherits(x, "SubTreeFileSystem")) {
+    stop("x must be a string or SubTreeFileSystem.")
+  }
   get_paths_and_filesystem(x, filesystem)
 }
 
-is_url <- function(x) is.string(x) && grepl("://", x)
-is_http_url <- function(x) is_url(x) && grepl("^http", x)
+is_url <- function(x) is_string(x) && grepl("://", x)
+is_http_url <- function(x) is_url(x) && startsWith(x, "http")
 are_urls <- function(x) if (!is.character(x)) FALSE else grepl("://", x)
 
 #' @usage NULL
@@ -538,7 +542,7 @@ default_s3_options <- list(
 #'
 #' @export
 s3_bucket <- function(bucket, ...) {
-  assert_that(is.string(bucket))
+  check_string(bucket)
   args <- list2(...)
 
   # If user specifies args, they must specify region as arg, env var, or config
@@ -573,7 +577,7 @@ s3_bucket <- function(bucket, ...) {
 #' bucket <- gs_bucket("arrow-datasets")
 #' @export
 gs_bucket <- function(bucket, ...) {
-  assert_that(is.string(bucket))
+  check_string(bucket)
   args <- list2(...)
 
   fs <- exec(GcsFileSystem$create, !!!args)
@@ -755,7 +759,7 @@ AzureFileSystem$create <- function(account_name, ...) {
 #' )
 #' @export
 az_container <- function(container_path, ...) {
-  assert_that(is.string(container_path))
+  check_string(container_path)
   args <- list2(...)
 
   fs <- exec(AzureFileSystem$create, !!!args)
@@ -798,7 +802,7 @@ SubTreeFileSystem$create <- function(base_path, base_fs = NULL) {
 #' @export
 `$.SubTreeFileSystem` <- function(x, name, ...) {
   # This is to allow delegating methods/properties to the base_fs
-  assert_that(is.string(name))
+  check_string(name)
   if (name %in% ls(envir = x)) {
     get(name, x)
   } else if (name %in% ls(envir = x$base_fs)) {
diff --git a/r/R/io.R b/r/R/io.R
index 06bd47c0a882..66919636abd8 100644
--- a/r/R/io.R
+++ b/r/R/io.R
@@ -246,7 +246,7 @@ make_readable_file <- function(file, mmap = TRUE, random_access = TRUE) {
     # file names with trailing slashes, so we need to remove it here.
     path <- sub("/$", "", file$base_path)
     file <- filesystem$OpenInputFile(path)
-  } else if (is.string(file)) {
+  } else if (is_string(file)) {
     # if this is a HTTP URL, we need a local copy to pass to FileSystem$from_uri
     if (random_access && is_http_url(file)) {
       tf <- tempfile()
@@ -303,7 +303,7 @@ make_output_stream <- function(x) {
     fs_and_path <- FileSystem$from_uri(x)
     fs_and_path$fs$OpenOutputStream(fs_and_path$path)
   } else {
-    assert_that(is.string(x))
+    check_string(x)
     FileOutputStream$create(x)
   }
 }
@@ -312,7 +312,7 @@ detect_compression <- function(path) {
   if (inherits(path, "SubTreeFileSystem")) {
     path <- path$base_path
   }
-  if (!is.string(path)) {
+  if (!is_string(path)) {
     return("uncompressed")
   }
 
diff --git a/r/R/parquet.R b/r/R/parquet.R
index 6415e36b03ce..c5d47f7765f3 100644
--- a/r/R/parquet.R
+++ b/r/R/parquet.R
@@ -251,7 +251,7 @@ make_valid_parquet_version <- function(version, valid_versions = valid_parquet_v
     version <- format(version, nsmall = 1)
   }
 
-  if (!is.string(version)) {
+  if (!is_string(version)) {
     stop(
       "`version` must be one of ",
       oxford_paste(names(valid_versions), "or"),
@@ -313,7 +313,7 @@ ParquetWriterPropertiesBuilder <- R6Class(
     },
     set_compression = function(column_names, compression) {
       compression <- compression_from_name(compression)
-      assert_that(is.integer(compression))
+      assert_is(compression, "integer")
       private$.set(
         column_names,
         compression,
@@ -330,7 +330,7 @@ ParquetWriterPropertiesBuilder <- R6Class(
       )
     },
     set_dictionary = function(column_names, use_dictionary) {
-      assert_that(is.logical(use_dictionary))
+      check_logical(use_dictionary)
       private$.set(
         column_names,
         use_dictionary,
@@ -338,7 +338,7 @@ ParquetWriterPropertiesBuilder <- R6Class(
       )
     },
     set_write_statistics = function(column_names, write_statistics) {
-      assert_that(is.logical(write_statistics))
+      check_logical(write_statistics)
       private$.set(
         column_names,
         write_statistics,
diff --git a/r/R/query-engine.R b/r/R/query-engine.R
index f1cbebb72a89..6c823879bd5f 100644
--- a/r/R/query-engine.R
+++ b/r/R/query-engine.R
@@ -372,7 +372,7 @@ head.ExecPlanReader <- function(x, n = 6L, ...) {
 }
 
 do_exec_plan_substrait <- function(substrait_plan) {
-  if (is.string(substrait_plan)) {
+  if (is_string(substrait_plan)) {
     substrait_plan <- substrait__internal__SubstraitFromJSON(substrait_plan)
   } else if (is.raw(substrait_plan)) {
     substrait_plan <- buffer(substrait_plan)
diff --git a/r/R/record-batch-reader.R b/r/R/record-batch-reader.R
index d979a3f9fe92..ad6b3c612fcd 100644
--- a/r/R/record-batch-reader.R
+++ b/r/R/record-batch-reader.R
@@ -134,10 +134,8 @@ as.data.frame.RecordBatchReader <- function(x, row.names = NULL, optional = FALS
 
 #' @export
 head.RecordBatchReader <- function(x, n = 6L, ...) {
-  assert_is(n, c("numeric", "integer"))
-  assert_that(length(n) == 1)
   # Negative n requires knowing nrow(x), which requires consuming the whole RBR
-  assert_that(n >= 0)
+  check_number_decimal(n, min = 0)
   if (!is.integer(n)) {
     n <- floor(n)
   }
@@ -251,7 +249,7 @@ as_record_batch_reader.Dataset <- function(x, ...) {
 #' @rdname as_record_batch_reader
 #' @export
 as_record_batch_reader.function <- function(x, ..., schema) {
-  assert_that(inherits(schema, "Schema"))
+  assert_is(schema, "Schema")
   RecordBatchReader__from_function(x, schema)
 }
 
diff --git a/r/R/record-batch-writer.R b/r/R/record-batch-writer.R
index eb96592a0f5b..49766b7a6000 100644
--- a/r/R/record-batch-writer.R
+++ b/r/R/record-batch-writer.R
@@ -118,7 +118,7 @@ RecordBatchWriter <- R6Class(
 #' @export
 RecordBatchStreamWriter <- R6Class("RecordBatchStreamWriter", inherit = RecordBatchWriter)
 RecordBatchStreamWriter$create <- function(sink, schema, use_legacy_format = NULL, metadata_version = NULL) {
-  if (is.string(sink)) {
+  if (is_string(sink)) {
     stop(
       "RecordBatchStreamWriter$create() requires an Arrow InputStream. ",
       "Try providing FileOutputStream$create(",
@@ -144,7 +144,7 @@ RecordBatchStreamWriter$create <- function(sink, schema, use_legacy_format = NUL
 #' @export
 RecordBatchFileWriter <- R6Class("RecordBatchFileWriter", inherit = RecordBatchStreamWriter)
 RecordBatchFileWriter$create <- function(sink, schema, use_legacy_format = NULL, metadata_version = NULL) {
-  if (is.string(sink)) {
+  if (is_string(sink)) {
     stop(
       "RecordBatchFileWriter$create() requires an Arrow InputStream. ",
       "Try providing FileOutputStream$create(",
diff --git a/r/R/record-batch.R b/r/R/record-batch.R
index f175b70ac9d9..c9b88768a1d8 100644
--- a/r/R/record-batch.R
+++ b/r/R/record-batch.R
@@ -91,7 +91,7 @@ RecordBatch <- R6Class(
       inherits(other, "RecordBatch") && RecordBatch__Equals(self, other, isTRUE(check_metadata))
     },
     GetColumnByName = function(name) {
-      assert_that(is.string(name))
+      check_string(name)
       RecordBatch__GetColumnByName(self, name)
     },
     SelectColumns = function(indices) RecordBatch__SelectColumns(self, indices),
@@ -119,7 +119,9 @@ RecordBatch <- R6Class(
     },
     cast = function(target_schema, safe = TRUE, ..., options = cast_options(safe, ...)) {
       assert_is(target_schema, "Schema")
-      assert_that(identical(self$schema$names, target_schema$names), msg = "incompatible schemas")
+      if (!identical(self$schema$names, target_schema$names)) {
+        stop("incompatible schemas", call. = FALSE)
+      }
       RecordBatch__cast(self, target_schema, options)
     },
     export_to_c = function(array_ptr, schema_ptr) {
diff --git a/r/R/schema.R b/r/R/schema.R
index a02753fac55a..7307f5b8d9f9 100644
--- a/r/R/schema.R
+++ b/r/R/schema.R
@@ -227,7 +227,7 @@ prepare_key_value_metadata <- function(metadata) {
 
 # Alternative to Schema__ToString that doesn't print metadata
 print_schema_fields <- function(s, truncate = FALSE, max_fields = 20L) {
-  assert_that(max_fields > 0)
+  check_number_whole(max_fields, min = 1)
   num_fields <- length(s$fields)
   if (truncate && num_fields > max_fields) {
     fields_out <- paste(map_chr(s$fields[seq_len(max_fields)], ~ .$ToString()), collapse = "\n")
@@ -318,7 +318,9 @@ length.Schema <- function(x) x$num_fields
 
 #' @export
 `[[<-.Schema` <- function(x, i, value) {
-  assert_that(length(i) == 1)
+  if (length(i) != 1) {
+    stop("`i` must have length 1.", call. = FALSE)
+  }
   if (is.character(i)) {
     field_names <- names(x)
     if (anyDuplicated(field_names)) {
@@ -333,7 +335,7 @@ length.Schema <- function(x) x$num_fields
     # No match means we're adding to the end
     i <- match(i, field_names, nomatch = length(field_names) + 1L)
   } else {
-    assert_that(is.numeric(i), !is.na(i), i > 0)
+    check_number_whole(i, min = 1, allow_na = FALSE)
     # If i is numeric and we have a type,
     # we need to grab the existing field name for the new one
     if (!is.null(value) && !inherits(value, "Field")) {
@@ -384,7 +386,7 @@ length.Schema <- function(x) x$num_fields
 
 #' @export
 `$.Schema` <- function(x, name, ...) {
-  assert_that(is.string(name))
+  check_string(name)
   if (name %in% ls(x)) {
     get(name, x)
   } else {
diff --git a/r/R/table.R b/r/R/table.R
index 4e8662e6e719..a25a6724014a 100644
--- a/r/R/table.R
+++ b/r/R/table.R
@@ -85,8 +85,7 @@ Table <- R6Class(
     nbytes = function() Table__ReferencedBufferSize(self),
     RenameColumns = function(value) Table__RenameColumns(self, value),
     GetColumnByName = function(name) {
-      assert_is(name, "character")
-      assert_that(length(name) == 1)
+      check_string(name)
       Table__GetColumnByName(self, name)
     },
     RemoveColumn = function(i) Table__RemoveColumn(self, i),
@@ -102,7 +101,9 @@ Table <- R6Class(
     },
     cast = function(target_schema, safe = TRUE, ..., options = cast_options(safe, ...)) {
       assert_is(target_schema, "Schema")
-      assert_that(identical(self$schema$names, target_schema$names), msg = "incompatible schemas")
+      if (!identical(self$schema$names, target_schema$names)) {
+        stop("Incompatible schemas", call. = FALSE)
+      }
       Table__cast(self, target_schema, options)
     },
     SelectColumns = function(indices) Table__SelectColumns(self, indices),
diff --git a/r/R/type.R b/r/R/type.R
index 14a4c8f1d2b6..01440623e10f 100644
--- a/r/R/type.R
+++ b/r/R/type.R
@@ -608,7 +608,7 @@ timestamp <- function(unit = c("s", "ms", "us", "ns"), timezone = "") {
     unit <- match.arg(unit)
   }
   unit <- make_valid_time_unit(unit, c(valid_time64_units, valid_time32_units))
-  assert_that(is.string(timezone))
+  check_string(timezone)
   Timestamp__initialize(unit, timezone)
 }
 
@@ -775,16 +775,14 @@ as_type <- function(type, name = "type") {
   if (identical(type, double())) {
     type <- float64()
   }
-  if (!inherits(type, "DataType")) {
-    stop(name, " must be a DataType, not ", class(type), call. = FALSE)
-  }
+  assert_is(type, "DataType", arg = name, call = NULL)
   type
 }
 
 canonical_type_str <- function(type_str) {
   # canonicalizes data type strings, converting data type function names and
   # aliases to match the strings returned by DataType$ToString()
-  assert_that(is.string(type_str))
+  check_string(type_str)
   if (grepl("[([<]", type_str)) {
     stop("Cannot interpret string representations of data types that have parameters", call. = FALSE)
   }
diff --git a/r/R/udf.R b/r/R/udf.R
index ce7a911e0d97..24decdbf734d 100644
--- a/r/R/udf.R
+++ b/r/R/udf.R
@@ -72,7 +72,7 @@
 #'   head()
 #'
 register_scalar_function <- function(name, fun, in_type, out_type, auto_convert = FALSE) {
-  assert_that(is.string(name))
+  check_string(name)
 
   scalar_function <- arrow_scalar_function(
     fun,
@@ -99,7 +99,9 @@ register_scalar_function <- function(name, fun, in_type, out_type, auto_convert
 }
 
 arrow_scalar_function <- function(fun, in_type, out_type, auto_convert = FALSE) {
-  assert_that(is.function(fun))
+  if (!is.function(fun)) {
+    stop("`fun` must be a function.")
+  }
 
   # Create a small wrapper function that is easier to call from C++.
   # TODO(ARROW-17148): This wrapper could be implemented in C/C++ to
diff --git a/r/R/util.R b/r/R/util.R
index dc48c5fb8ff8..899f565bd4a9 100644
--- a/r/R/util.R
+++ b/r/R/util.R
@@ -30,14 +30,135 @@ oxford_paste <- function(x, conjunction = "and", quote = TRUE, quote_symbol = '"
   }
 }
 
-assert_is <- function(object, class) {
-  msg <- paste(substitute(object), "must be a", oxford_paste(class, "or"))
-  assert_that(inherits(object, class), msg = msg)
+check_null <- function(x, msg = NULL, call = rlang::caller_env()) {
+  if (!is.null(x)) {
+    if (is.null(msg)) {
+      rlang::stop_input_type(
+        x,
+        what = "NULL",
+        call = call
+      )
+    } else {
+      abort(msg, call = call)
+    }
+  }
+}
+
+# copied from https://github.com/r-lib/rlang/blob/main/R/standalone-types-check.R
+# they are not licensed.
+
+check_character <- function(
+  x,
+  ...,
+  allow_na = TRUE,
+  allow_null = FALSE,
+  arg = caller_arg(x),
+  call = caller_env()
+) {
+  if (!missing(x)) {
+    if (is_character(x)) {
+      if (!allow_na && any(is.na(x))) {
+        abort(
+          sprintf("`%s` can't contain NA values.", arg),
+          arg = arg,
+          call = call
+        )
+      }
+
+      return(invisible(NULL))
+    }
+
+    if (allow_null && is_null(x)) {
+      return(invisible(NULL))
+    }
+  }
+
+  stop_input_type(
+    x,
+    "a character vector",
+    ...,
+    allow_null = allow_null,
+    arg = arg,
+    call = call
+  )
+}
+
+check_logical <- function(
+  x,
+  ...,
+  allow_na = TRUE,
+  allow_null = FALSE,
+  arg = caller_arg(x),
+  call = caller_env()
+) {
+  if (!missing(x)) {
+    if (is_logical(x)) {
+      if (!allow_na && any(is.na(x))) {
+        abort(
+          sprintf("`%s` can't contain NA values.", arg),
+          arg = arg,
+          call = call
+        )
+      }
+      return(invisible(NULL))
+    }
+    if (allow_null && is_null(x)) {
+      return(invisible(NULL))
+    }
+  }
+
+  stop_input_type(
+    x,
+    "a logical vector",
+    ...,
+    allow_na = FALSE,
+    allow_null = allow_null,
+    arg = arg,
+    call = call
+  )
+}
+
+assert_is <- function(
+  object,
+  class,
+  call = caller_env(),
+  arg = caller_arg(object),
+  allow_null = FALSE
+) {
+  if (is.null(object) && allow_null) {
+    return(invisible(NULL))
+  }
+
+  if (!inherits(object, class)) {
+    # recode for friendlier error message
+    friendly_types <- c(
+      "factor" = "a factor",
+      "Array" = "an Array",
+      "RandomAccessFile" = "a RandomAccessFile",
+      "DataType" = "a DataType",
+      "data.frame" = "a data frame"
+    )
+    class <- vctrs::vec_replace_values(
+      class,
+      from = names(friendly_types),
+      to = unname(friendly_types)
+    )
+    rlang::stop_input_type(
+      x = object,
+      what = class,
+      call = call,
+      arg = arg,
+      allow_null = allow_null,
+      class = "assertError"
+    )
+  }
 }
 
 assert_is_list_of <- function(object, class) {
   msg <- paste(substitute(object), "must be a list of", oxford_paste(class, "or"))
-  assert_that(is_list_of(object, class), msg = msg)
+  if (!is_list_of(object, class)) {
+    stop(msg, call. = FALSE)
+  }
 }
 
 is_list_of <- function(object, class) {
diff --git a/r/extra-tests/test-read-files.R b/r/extra-tests/test-read-files.R
index ced366d2f571..8e25451286cf 100644
--- a/r/extra-tests/test-read-files.R
+++ b/r/extra-tests/test-read-files.R
@@ -38,10 +38,7 @@ pq_file <- "files/ex_data.parquet"
 
 test_that("Can read the file (parquet)", {
   # We can read with no error, we assert metadata below
-  expect_error(
-    df <- read_parquet(pq_file),
-    NA
-  )
+  expect_no_error(df <- read_parquet(pq_file))
 })
 
 ### Parquet
@@ -83,10 +80,7 @@ for (comp in c("lz4", "uncompressed", "zstd")) {
 
   test_that(paste0("Can read the file (feather ", comp, ")"), {
     # We can read with no error, we assert metadata below
-    expect_error(
-      df <- read_feather(feather_file),
-      NA
-    )
+    expect_no_error(df <- read_feather(feather_file))
   })
 
   test_that(paste0("Can see the metadata (feather ", comp, ")"), {
diff --git a/r/man/arrow_not_supported.Rd b/r/man/arrow_not_supported.Rd
index be6a001fa1fa..bbcc82e4f4da 100644
--- a/r/man/arrow_not_supported.Rd
+++ b/r/man/arrow_not_supported.Rd
@@ -49,7 +49,7 @@ data into R. If you have an \code{arrow_not_supported()} error with a \code{>}
 suggestion, when the error is ultimately raised by \code{try_error_dplyr()},
 \verb{Call collect() first to pull data into R} won't be the only suggestion.
 
-You can still use \code{match.arg()} and \code{assert_that()} for simple input
+You can still use \code{match.arg()} and \code{assert_is()} for simple input
 validation inside of the function bindings. \code{arrow_eval()} will catch their
 errors and re-raise them as \code{validation_error}.
 }
diff --git a/r/tests/testthat/_snaps/csv.md b/r/tests/testthat/_snaps/csv.md
new file mode 100644
index 000000000000..4f50baab4443
--- /dev/null
+++ b/r/tests/testthat/_snaps/csv.md
@@ -0,0 +1,93 @@
+# Writing a CSV errors when unsupported (yet) readr args are used
+
+    Code
+      write_csv_arrow(tbl, csv_file, append = FALSE)
+    Condition
+      Error in `write_csv_arrow()`:
+      ! Arguments not yet supported in Arrow
+      * `...` must be empty.
+      x Problematic argument:
+      * append = FALSE
+
+---
+
+    Code
+      write_csv_arrow(tbl, csv_file, quote = "all")
+    Condition
+      Error in `write_csv_arrow()`:
+      ! Arguments not yet supported in Arrow
+      * `...` must be empty.
+      x Problematic argument:
+      * quote = "all"
+
+---
+
+    Code
+      write_csv_arrow(tbl, csv_file, escape = "double")
+    Condition
+      Error in `write_csv_arrow()`:
+      ! Arguments not yet supported in Arrow
+      * `...` must be empty.
+      x Problematic argument:
+      * escape = "double"
+
+---
+
+    Code
+      write_csv_arrow(tbl, csv_file, eol = "\n")
+    Condition
+      Error in `write_csv_arrow()`:
+      ! Arguments not yet supported in Arrow
+      * `...` must be empty.
+      x Problematic argument:
+      * eol = "\n"
+
+---
+
+    Code
+      write_csv_arrow(tbl, csv_file, num_threads = 8)
+    Condition
+      Error in `write_csv_arrow()`:
+      ! Arguments not yet supported in Arrow
+      * `...` must be empty.
+      x Problematic argument:
+      * num_threads = 8
+
+---
+
+    Code
+      write_csv_arrow(tbl, csv_file, progress = FALSE)
+    Condition
+      Error in `write_csv_arrow()`:
+      ! Arguments not yet supported in Arrow
+      * `...` must be empty.
+      x Problematic argument:
+      * progress = FALSE
+
+---
+
+    Code
+      write_csv_arrow(tbl, csv_file, append = FALSE, eol = "\n")
+    Condition
+      Error in `write_csv_arrow()`:
+      ! Arguments not yet supported in Arrow
+      * `...` must be empty.
+      x Problematic arguments:
+      * append = FALSE
+      * eol = "\n"
+
+---
+
+    Code
+      write_csv_arrow(tbl, csv_file, append = FALSE, quote = "all", escape = "double",
+        eol = "\n")
+    Condition
+      Error in `write_csv_arrow()`:
+      ! Arguments not yet supported in Arrow
+      * `...` must be empty.
+      x Problematic arguments:
+      * append = FALSE
+      * quote = "all"
+      * escape = "double"
+      * eol = "\n"
+
diff --git a/r/tests/testthat/_snaps/udf.md b/r/tests/testthat/_snaps/udf.md
index 89506a7fbc23..27d17e11db4b 100644
--- a/r/tests/testthat/_snaps/udf.md
+++ b/r/tests/testthat/_snaps/udf.md
@@ -1,4 +1,4 @@
 # arrow_scalar_function() works
 
-    fun is not a function
+    `fun` must be a function.
 
diff --git a/r/tests/testthat/test-Array.R b/r/tests/testthat/test-Array.R
index b5233eb30360..89b186647d0c 100644
--- a/r/tests/testthat/test-Array.R
+++ b/r/tests/testthat/test-Array.R
@@ -97,7 +97,7 @@ test_that("Slice() and RangeEquals()", {
 
   y <- x$Slice(10)
   expect_equal(y$type, int32())
-  expect_equal(length(y), 15L)
+  expect_length(y, 15L)
   expect_as_vector(y, c(101:110, 201:205))
   expect_true(x$RangeEquals(y, 10, 24))
   expect_false(x$RangeEquals(y, 9, 23))
@@ -122,7 +122,7 @@ test_that("Slice() and RangeEquals()", {
   expect_error(x$Slice(10, -1), "Slice 'length' cannot be negative")
   expect_error(x$Slice(-1, 10), "Slice 'offset' cannot be negative")
 
-  expect_warning(x$Slice(10, 15), NA)
+  expect_no_warning(x$Slice(10, 15))
   expect_warning(
     overslice <- x$Slice(10, 16),
     "Slice 'length' greater than available length"
@@ -130,7 +130,7 @@ test_that("Slice() and RangeEquals()", {
   expect_equal(length(overslice), 15)
   expect_warning(z$Slice(2, 10), "Slice 'length' greater than available length")
 
-  expect_error(x$RangeEquals(10, 24, 0), 'other must be a "Array"')
+  expect_error(x$RangeEquals(10, 24, 0), "`other` must be an Array")
   expect_error(x$RangeEquals(y, NA, 24), "'start_idx' cannot be NA")
   expect_error(x$RangeEquals(y, 10, NA), "'end_idx' cannot be NA")
   expect_error(x$RangeEquals(y, 10, 24, NA), "'other_start_idx' cannot be NA")
@@ -428,7 +428,7 @@ test_that("integer types cast safety (ARROW-3741, ARROW-5541)", {
   a <- arrow_array(-(1:10))
   for (type in uint_types) {
     expect_error(a$cast(type), regexp = "Integer value -1 not in range")
-    expect_error(a$cast(type, safe = FALSE), NA)
+    expect_no_error(a$cast(type, safe = FALSE))
   }
 })
 
@@ -452,7 +452,7 @@ test_that("cast to half float works", {
 
 test_that("cast input validation", {
   a <- arrow_array(1:4)
-  expect_error(a$cast("not a type"), "type must be a DataType, not character")
+  expect_error(a$cast("not a type"), "`type` must be a DataType, not")
 })
 
 test_that("arrow_array() supports the type= argument. conversion from INTSXP and int64 to all int types", {
@@ -473,7 +473,7 @@ test_that("arrow_array() supports the type= argument. conversion from INTSXP and
   # Input validation
   expect_error(
     arrow_array(5, type = "not a type"),
-    "type must be a DataType, not character"
+    "`type` must be a DataType, not"
   )
 })
 
@@ -869,7 +869,7 @@ test_that("Handling string data with embedded nuls", {
   array_with_nul <- arrow_array(raws)$cast(utf8())
 
   # no error on conversion, because altrep laziness
-  v <- expect_error(as.vector(array_with_nul), NA)
+  expect_no_error(v <- as.vector(array_with_nul))
 
   # attempting materialization -> error
 
@@ -946,12 +946,12 @@ test_that("Array$View() (ARROW-6542)", {
   expect_equal(length(b), 3L)
 
   # Input validation
-  expect_error(a$View("not a type"), "type must be a DataType, not character")
+  expect_error(a$View("not a type"), "`type` must be a DataType, not")
 })
 
 test_that("Array$Validate()", {
   a <- arrow_array(1:10)
-  expect_error(a$Validate(), NA)
+  expect_no_error(a$Validate())
 })
 
 test_that("is.Array", {
diff --git a/r/tests/testthat/test-RecordBatch.R b/r/tests/testthat/test-RecordBatch.R
index dbb21c06568e..a3402eb5c15c 100644
--- a/r/tests/testthat/test-RecordBatch.R
+++ b/r/tests/testthat/test-RecordBatch.R
@@ -161,9 +161,9 @@ test_that("[[ and $ on RecordBatch", {
   expect_null(batch$qwerty)
   expect_null(batch[["asdf"]])
   expect_error(batch[[c(4, 3)]])
-  expect_error(batch[[NA]], "'i' must be character or numeric, not logical")
-  expect_error(batch[[NULL]], "'i' must be character or numeric, not NULL")
-  expect_error(batch[[c("asdf", "jkl;")]], "name is not a string", fixed = TRUE)
+  expect_error(batch[[NA]], "`i` must be character or numeric, not `NA`")
+  expect_error(batch[[NULL]], "`i` must be character or numeric, not `NULL`")
+  expect_error(batch[[c("asdf", "jkl;")]], "`name` must be a single string", fixed = TRUE)
 })
 
 test_that("[[<- assignment", {
@@ -224,12 +224,12 @@ test_that("[[<- assignment", {
   expect_as_vector(batch$array, 10:1)
 
   # nonsense indexes
-  expect_error(batch[[NA]] <- letters[10:1], "'i' must be character or numeric, not logical")
-  expect_error(batch[[NULL]] <- letters[10:1], "'i' must be character or numeric, not NULL")
-  expect_error(batch[[NA_integer_]] <- letters[10:1], "!is.na(i) is not TRUE", fixed = TRUE)
-  expect_error(batch[[NA_real_]] <- letters[10:1], "!is.na(i) is not TRUE", fixed = TRUE)
-  expect_error(batch[[NA_character_]] <- letters[10:1], "!is.na(i) is not TRUE", fixed = TRUE)
-  expect_error(batch[[c(1, 4)]] <- letters[10:1], "length(i) not equal to 1", fixed = TRUE)
+  expect_error(batch[[NA]] <- letters[10:1], "`i` must be character or numeric, not `NA`")
+  expect_error(batch[[NULL]] <- letters[10:1], "`i` must be character or numeric, not `NULL`")
+  expect_error(batch[[NA_integer_]] <- letters[10:1], "`i` must be", fixed = TRUE)
+  expect_error(batch[[NA_real_]] <- letters[10:1], "`i` must be", fixed = TRUE)
+  expect_error(batch[[NA_character_]] <- letters[10:1], "`i` must be", fixed = TRUE)
+  expect_error(batch[[c(1, 4)]] <- letters[10:1], "`i` must be", fixed = TRUE)
 })
 
 test_that("head and tail on RecordBatch", {
diff --git a/r/tests/testthat/test-Table.R b/r/tests/testthat/test-Table.R
index 12ef25975c70..5aed6ca02b5f 100644
--- a/r/tests/testthat/test-Table.R
+++ b/r/tests/testthat/test-Table.R
@@ -100,9 +100,9 @@ test_that("[, [[, $ for Table", {
   expect_equal_data_frame(tab[-3], tbl[-3])
 
   expect_error(tab[[c(4, 3)]])
-  expect_error(tab[[NA]], "'i' must be character or numeric, not logical")
-  expect_error(tab[[NULL]], "'i' must be character or numeric, not NULL")
-  expect_error(tab[[c("asdf", "jkl;")]], "length(name) not equal to 1", fixed = TRUE)
+  expect_error(tab[[NA]], "`i` must be character or numeric, not `NA`")
+  expect_error(tab[[NULL]], "`i` must be character or numeric, not `NULL`")
+  expect_error(tab[[c("asdf", "jkl;")]], "`name` must be", fixed = TRUE)
   expect_error(tab[-3:3], "Invalid column index")
   expect_error(tab[1000], "Invalid column index")
   expect_error(tab[1:1000], "Invalid column index")
@@ -167,12 +167,16 @@ test_that("[[<- assignment", {
   expect_as_vector(tab$chunked, 1:10)
 
   # nonsense indexes
-  expect_error(tab[[NA]] <- letters[10:1], "'i' must be character or numeric, not logical")
-  expect_error(tab[[NULL]] <- letters[10:1], "'i' must be character or numeric, not NULL")
-  expect_error(tab[[NA_integer_]] <- letters[10:1], "!is.na(i) is not TRUE", fixed = TRUE)
-  expect_error(tab[[NA_real_]] <- letters[10:1], "!is.na(i) is not TRUE", fixed = TRUE)
-  expect_error(tab[[NA_character_]] <- letters[10:1], "!is.na(i) is not TRUE", fixed = TRUE)
-  expect_error(tab[[c(1, 4)]] <- letters[10:1], "length(i) not equal to 1", fixed = TRUE)
+  expect_error(tab[[NA]] <- letters[10:1], "`i` must be character or numeric, not `NA`")
+  expect_error(tab[[NULL]] <- letters[10:1], "`i` must be character or numeric, not `NULL`")
+  expect_error(
+    tab[[NA_integer_]] <- letters[10:1],
+    "`i` must be character or numeric, not an integer `NA`.",
+    fixed = TRUE
+  )
+  expect_error(tab[[NA_real_]] <- letters[10:1], "`i` must be length 1 and not NA", fixed = TRUE)
+  expect_error(tab[[NA_character_]] <- letters[10:1], "`i` must be length 1 and not NA", fixed = TRUE)
+  expect_error(tab[[c(1, 4)]] <- letters[10:1], "`i` must be length 1 and not NA", fixed = TRUE)
 })
 
 test_that("Table$Slice", {
diff --git a/r/tests/testthat/test-arrow.R b/r/tests/testthat/test-arrow.R
index c6ae27ac5229..ee3e1b92b479 100644
--- a/r/tests/testthat/test-arrow.R
+++ b/r/tests/testthat/test-arrow.R
@@ -25,17 +25,19 @@ test_that("Can't $new() an object with anything other than a pointer", {
 
 test_that("assert_is", {
   x <- 42
-  expect_true(assert_is(x, "numeric"))
-  expect_true(assert_is(x, c("numeric", "character")))
-  expect_error(assert_is(x, "factor"), 'x must be a "factor"')
+  expect_no_error(assert_is(x, "numeric"))
+  expect_no_error(assert_is(x, c("numeric", "character")))
+  expect_error(assert_is(x, "factor"), "`x` must be a factor")
   expect_error(
     assert_is(x, c("factor", "list")),
-    'x must be a "factor" or "list"'
+    "`x` must be a factor or list"
   )
   expect_error(
     assert_is(x, c("factor", "character", "list")),
-    'x must be a "factor", "character", or "list"'
+    "`x` must be a factor, character, or list"
   )
+  y <- NULL
+  expect_no_error(assert_is(y, class = "data.frame", allow_null = TRUE))
 })
 
 test_that("arrow gracefully fails to load objects from other sessions (ARROW-10071)", {
diff --git a/r/tests/testthat/test-chunked-array.R b/r/tests/testthat/test-chunked-array.R
index bcadaa889f90..bc50e3e3337f 100644
--- a/r/tests/testthat/test-chunked-array.R
+++ b/r/tests/testthat/test-chunked-array.R
@@ -86,7 +86,7 @@ test_that("ChunkedArray", {
   expect_error(x$Slice(10, -1), "Slice 'length' cannot be negative")
   expect_error(x$Slice(-1, 10), "Slice 'offset' cannot be negative")
 
-  expect_warning(x$Slice(10, 15), NA)
+  expect_no_warning(x$Slice(10, 15))
   expect_warning(
     overslice <- x$Slice(10, 16),
     "Slice 'length' greater than available length"
@@ -393,12 +393,12 @@ test_that("ChunkedArray$View() (ARROW-6542)", {
   expect_equal(length(b), 7L)
   expect_all_true(sapply(b$chunks, function(.x) .x$type == float32()))
   # Input validation
-  expect_error(a$View("not a type"), "type must be a DataType, not character")
+  expect_error(a$View("not a type"), "`type` must be a DataType, not")
 })
 
 test_that("ChunkedArray$Validate()", {
   a <- ChunkedArray$create(1:10)
-  expect_error(a$Validate(), NA)
+  expect_no_error(a$Validate())
 })
 
 test_that("[ ChunkedArray", {
@@ -488,7 +488,7 @@ test_that("Handling string data with embedded nuls", {
   ))
   chunked_array_with_nul <- ChunkedArray$create(raws)$cast(utf8())
 
-  v <- expect_error(as.vector(chunked_array_with_nul), NA)
+  expect_no_error(v <- as.vector(chunked_array_with_nul))
 
   expect_error(
     v[],
@@ -500,7 +500,7 @@ test_that("Handling string data with embedded nuls", {
   )
 
   withr::with_options(list(arrow.skip_nul = TRUE), {
-    v <- expect_warning(as.vector(chunked_array_with_nul), NA)
+    expect_no_warning(v <- as.vector(chunked_array_with_nul))
     expect_warning(
       expect_identical(v[3], "man"),
       "Stripping '\\0' (nul) from character vector",
diff --git a/r/tests/testthat/test-compute-aggregate.R b/r/tests/testthat/test-compute-aggregate.R
index dd79682361bf..dd69ccff0429 100644
--- a/r/tests/testthat/test-compute-aggregate.R
+++ b/r/tests/testthat/test-compute-aggregate.R
@@ -20,9 +20,9 @@ test_that("list_compute_functions", {
   expect_false(all(grepl("min", allfuncs)))
   justmins <- list_compute_functions("^min")
   expect_true(length(justmins) > 0)
-  expect_all_true(grepl("min", justmins))
+  expect_match(justmins, "min")
   no_hash_funcs <- list_compute_functions("^hash")
-  expect_true(length(no_hash_funcs) == 0)
+  expect_length(no_hash_funcs, 0)
 })
 
 test_that("sum.Array", {
diff --git a/r/tests/testthat/test-compute-vector.R b/r/tests/testthat/test-compute-vector.R
index bfc42b432f3f..9a2d129c2f25 100644
--- a/r/tests/testthat/test-compute-vector.R
+++ b/r/tests/testthat/test-compute-vector.R
@@ -135,9 +135,8 @@ test_that("call_function validation", {
     call_function("filter", Array$create(1:4), Array$create(c(TRUE, FALSE, TRUE)), options = list(keep_na = TRUE)),
     "Arguments for execution of vector kernel function 'array_filter' must all be the same length"
   )
-  expect_error(
-    call_function("filter", record_batch(a = 1:3), Array$create(c(TRUE, FALSE, TRUE)), options = list(keep_na = TRUE)),
-    NA
+  expect_no_error(
+    call_function("filter", record_batch(a = 1:3), Array$create(c(TRUE, FALSE, TRUE)), options = list(keep_na = TRUE))
   )
   expect_error(
     call_function("filter", options = list(keep_na = TRUE)),
diff --git a/r/tests/testthat/test-csv.R b/r/tests/testthat/test-csv.R
index 8fb11c2a5e31..09ae1ef1706e 100644
--- a/r/tests/testthat/test-csv.R
+++ b/r/tests/testthat/test-csv.R
@@ -430,7 +430,7 @@ test_that("Write a CSV file with invalid input type", {
 test_that("Write a CSV file with invalid batch size", {
   expect_error(
     write_csv_arrow(tbl_no_dates, csv_file, batch_size = -1),
-    regexp = "batch_size not greater than 0"
+    regexp = "`batch_size` must be"
   )
 })
 
@@ -440,7 +440,7 @@ test_that("Write a CSV with custom NA value", {
   expect_identical(tbl_out1, tbl_no_dates)
 
   csv_contents <- readLines(csv_file)
-  expect_true(any(grepl("NULL_VALUE", csv_contents)))
+  expect_match(csv_contents, "NULL_VALUE", all = FALSE)
 
   tbl_in1 <- read_csv_arrow(csv_file, na = "NULL_VALUE")
   expect_identical(tbl_in1, tbl_no_dates)
@@ -448,7 +448,7 @@ test_that("Write a CSV with custom NA value", {
   # Also can use null_value in CsvWriteOptions
   tbl_out1 <- write_csv_arrow(tbl_no_dates, csv_file, write_options = csv_write_options(null_string = "another_null"))
   csv_contents <- readLines(csv_file)
-  expect_true(any(grepl("another_null", csv_contents)))
+  expect_match(csv_contents, "another_null", all = FALSE)
 
   tbl_in1 <- read_csv_arrow(csv_file, na = "another_null")
   expect_identical(tbl_in1, tbl_no_dates)
@@ -458,7 +458,7 @@ test_that("Write a CSV with custom NA value", {
   expect_true(file.exists(csv_file))
 
   csv_contents <- readLines(csv_file)
-  expect_true(any(grepl(",,", csv_contents)))
+  expect_match(csv_contents, ",,", all = FALSE)
 
   tbl_in1 <- read_csv_arrow(csv_file)
   expect_identical(tbl_in1, tbl_no_dates)
@@ -490,35 +490,15 @@ test_that("time mapping work as expected (ARROW-13624)", {
 })
 
 test_that("Writing a CSV errors when unsupported (yet) readr args are used", {
-  expect_error(
-    write_csv_arrow(tbl, csv_file, append = FALSE),
-    "The following argument is not yet supported in Arrow: \"append\""
-  )
-  expect_error(
-    write_csv_arrow(tbl, csv_file, quote = "all"),
-    "The following argument is not yet supported in Arrow: \"quote\""
-  )
-  expect_error(
-    write_csv_arrow(tbl, csv_file, escape = "double"),
-    "The following argument is not yet supported in Arrow: \"escape\""
-  )
-  expect_error(
-    write_csv_arrow(tbl, csv_file, eol = "\n"),
-    "The following argument is not yet supported in Arrow: \"eol\""
-  )
-  expect_error(
-    write_csv_arrow(tbl, csv_file, num_threads = 8),
-    "The following argument is not yet supported in Arrow: \"num_threads\""
-  )
-  expect_error(
-    write_csv_arrow(tbl, csv_file, progress = FALSE),
-    "The following argument is not yet supported in Arrow: \"progress\""
-  )
-  expect_error(
-    write_csv_arrow(tbl, csv_file, append = FALSE, eol = "\n"),
-    "The following arguments are not yet supported in Arrow: \"append\" and \"eol\""
-  )
-  expect_error(
+  expect_snapshot(error = TRUE, write_csv_arrow(tbl, csv_file, append = FALSE))
+  expect_snapshot(error = TRUE, write_csv_arrow(tbl, csv_file, quote = "all"))
+  expect_snapshot(error = TRUE, write_csv_arrow(tbl, csv_file, escape = "double"))
+  expect_snapshot(error = TRUE, write_csv_arrow(tbl, csv_file, eol = "\n"))
+  expect_snapshot(error = TRUE, write_csv_arrow(tbl, csv_file, num_threads = 8))
+  expect_snapshot(error = TRUE, write_csv_arrow(tbl, csv_file, progress = FALSE), )
+  expect_snapshot(error = TRUE, write_csv_arrow(tbl, csv_file, append = FALSE, eol = "\n"))
+  expect_snapshot(
+    error = TRUE,
     write_csv_arrow(
       tbl,
       csv_file,
@@ -526,10 +506,6 @@ test_that("Writing a CSV errors when unsupported (yet) readr args are used", {
       quote = "all",
       escape = "double",
       eol = "\n"
-    ),
-    paste(
-      "The following arguments are not yet supported in Arrow: \"append\",",
-      "\"quote\", \"escape\", and \"eol\""
     )
   )
 })
diff --git a/r/tests/testthat/test-data-type.R b/r/tests/testthat/test-data-type.R
index fa2e5bcd6e8d..a868246f5edd 100644
--- a/r/tests/testthat/test-data-type.R
+++ b/r/tests/testthat/test-data-type.R
@@ -343,11 +343,11 @@ test_that("timestamp type input validation", {
   )
   expect_error(
     timestamp(timezone = 1231231),
-    "timezone is not a string"
+    "`timezone` must be"
   )
   expect_error(
     timestamp(timezone = c("not", "a", "timezone")),
-    "timezone is not a string"
+    "`timezone` must be"
   )
 })
 
@@ -469,8 +469,8 @@ test_that("DictionaryType validation", {
     dictionary(utf8(), int32()),
     "Dictionary index type should be .*integer, got string"
   )
-  expect_error(dictionary(4, utf8()), 'index_type must be a "DataType"')
-  expect_error(dictionary(int8(), "strings"), 'value_type must be a "DataType"')
+  expect_error(dictionary(4, utf8()), '`index_type` must be a DataType')
+  expect_error(dictionary(int8(), "strings"), '`value_type` must be a DataType')
 })
 
 test_that("decimal type and validation", {
diff --git a/r/tests/testthat/test-dataset-dplyr.R b/r/tests/testthat/test-dataset-dplyr.R
index 09ca0ef155b7..6504ace7e863 100644
--- a/r/tests/testthat/test-dataset-dplyr.R
+++ b/r/tests/testthat/test-dataset-dplyr.R
@@ -275,7 +275,7 @@ test_that("compute()/collect(as_data_frame=FALSE)", {
 
   expect_r6_class(tab5, "Table")
   # mutate() was evaluated
-  expect_true("negint" %in% names(tab5))
+  expect_contains(names(tab5), "negint")
 })
 
 test_that("head/tail on query on dataset", {
diff --git a/r/tests/testthat/test-dataset-write.R b/r/tests/testthat/test-dataset-write.R
index 8fed358dc372..1614a92d0d77 100644
--- a/r/tests/testthat/test-dataset-write.R
+++ b/r/tests/testthat/test-dataset-write.R
@@ -67,7 +67,7 @@ test_that("Writing a dataset: CSV->IPC", {
     as_data_frame = FALSE
   )
   # It shouldn't be there
-  expect_false("int" %in% names(first))
+  expect_disjoint(names(first), "int")
 })
 
 test_that("Writing a dataset: Parquet->IPC", {
@@ -221,9 +221,7 @@ test_that("Writing a dataset: no format specified", {
     list.files(dst_dir, pattern = "parquet"),
     "part-0.parquet"
   )
-  expect_true(
-    inherits(new_ds$format, "ParquetFileFormat")
-  )
+  expect_s3_class(new_ds$format, "ParquetFileFormat")
   expect_equal(
     new_ds |> collect(),
     example_data
@@ -581,9 +579,7 @@ test_that("max_rows_per_file = 0 does not trigger max_rows_per_group adjustment
 
   # max_rows_per_file = 0 means "no limit" and should not error
   dst_dir <- make_temp_dir()
-  expect_no_error(
-    write_dataset(df1, dst_dir, max_rows_per_file = 0L)
-  )
+  expect_no_error(write_dataset(df1, dst_dir, max_rows_per_file = 0L))
 })
 
 
diff --git a/r/tests/testthat/test-dataset.R b/r/tests/testthat/test-dataset.R
index a64ea4cc47f6..2c3135b5f3dc 100644
--- a/r/tests/testthat/test-dataset.R
+++ b/r/tests/testthat/test-dataset.R
@@ -690,16 +690,8 @@ test_that("scalar aggregates with many batches (ARROW-16904)", {
   ds <- open_dataset(tf)
   replicate(100, ds |> summarize(min(x)) |> pull())
 
-  expect_true(
-    all(
-      replicate(100, ds |> summarize(min(x)) |> pull() |> as.vector()) == 1
-    )
-  )
-  expect_true(
-    all(
-      replicate(100, ds |> summarize(max(x)) |> pull() |> as.vector()) == 100
-    )
-  )
+  expect_all_true(replicate(100, ds |> summarize(min(x)) |> pull() |> as.vector()) == 1)
+  expect_all_true(replicate(100, ds |> summarize(max(x)) |> pull() |> as.vector()) == 100)
 })
 
 test_that("streaming map_batches into an ExecPlan", {
@@ -1446,7 +1438,7 @@ test_that("FileSystemFactoryOptions input validation", {
       partitioning = "part",
       factory_options = list(partition_base_dir = 42)
     ),
-    "factory_options$partition_base_dir is not a string",
+    "`factory_options$partition_base_dir` must be a single string",
     fixed = TRUE
   )
   expect_error(
diff --git a/r/tests/testthat/test-dplyr-filter.R b/r/tests/testthat/test-dplyr-filter.R
index ad69b26be798..a5416d1ad476 100644
--- a/r/tests/testthat/test-dplyr-filter.R
+++ b/r/tests/testthat/test-dplyr-filter.R
@@ -287,27 +287,25 @@ test_that("filter environment scope", {
 })
 
 test_that("Filtering on a column that doesn't exist errors correctly", {
-  # expect_warning(., NA) because the usual behavior when it hits a filter
+  # expect_no_warning() because the usual behavior when it hits a filter
   # that it can't evaluate is to raise a warning, collect() to R, and retry
   # the filter. But we want this to error the first time because it's
   # a user error, not solvable by retrying in R
-  expect_warning(
+  expect_no_warning(
     expect_error(
       tbl |> record_batch() |> filter(not_a_col == 42) |> collect(),
       "object 'not_a_col' not found"
-    ),
-    NA
+    )
   )
 })
 
 test_that("Filtering on a non-existent column errors in the correct language", {
   with_language("fr", {
-    expect_warning(
+    expect_no_warning(
       expect_error(
         tbl |> record_batch() |> filter(not_a_col == 42) |> collect(),
         "objet 'not_a_col' introuvable"
-      ),
-      NA
+      )
     )
   })
 })
diff --git a/r/tests/testthat/test-dplyr-mutate.R b/r/tests/testthat/test-dplyr-mutate.R
index 3e116a1012ca..f918b0007165 100644
--- a/r/tests/testthat/test-dplyr-mutate.R
+++ b/r/tests/testthat/test-dplyr-mutate.R
@@ -514,16 +514,15 @@ test_that("handle bad expressions", {
   # that need to be forced to fail because they error ambiguously
 
   with_language("fr", {
-    # expect_warning(., NA) because the usual behavior when it hits a filter
+    # expect_no_warning() because the usual behavior when it hits a filter
     # that it can't evaluate is to raise a warning, collect() to R, and retry
     # the filter. But we want this to error the first time because it's
     # a user error, not solvable by retrying in R
-    expect_warning(
+    expect_no_warning(
       expect_error(
         Table$create(tbl) |> mutate(newvar = NOTAVAR + 2),
         "objet 'NOTAVAR' introuvable"
-      ),
-      NA
+      )
     )
   })
 })
diff --git a/r/tests/testthat/test-duckdb.R b/r/tests/testthat/test-duckdb.R
index c0e69026488d..8c5611f900e1 100644
--- a/r/tests/testthat/test-duckdb.R
+++ b/r/tests/testthat/test-duckdb.R
@@ -236,7 +236,7 @@ test_that("Joining, auto-cleanup enabled", {
       ".int"
     )
   )
-  expect_identical(dim(res), c(9L, 14L))
+  expect_shape(res, dim = c(9L, 14L))
 
   # clean up cleans up the tables
   expect_all_true(c(table_one_name, table_two_name) %in% duckdb::duckdb_list_arrow(con))
diff --git a/r/tests/testthat/test-expression.R b/r/tests/testthat/test-expression.R
index 2f40c621ec8a..f6a8e5a8be66 100644
--- a/r/tests/testthat/test-expression.R
+++ b/r/tests/testthat/test-expression.R
@@ -148,7 +148,7 @@ test_that("Nested field ref types", {
 
 test_that("Nested field from a non-field-ref (struct_field kernel)", {
   x <- Expression$scalar(tibble::tibble(a = 1, b = "two"))
-  expect_true(inherits(x$a, "Expression"))
+  expect_s3_class(x$a, "Expression")
   expect_equal(x$a$type(), float64())
   expect_error(x$c, "field 'c' not found in struct<a: double, b: string>")
 })
diff --git a/r/tests/testthat/test-feather.R b/r/tests/testthat/test-feather.R
index 188a562fe81b..6a2a4c2fdb14 100644
--- a/r/tests/testthat/test-feather.R
+++ b/r/tests/testthat/test-feather.R
@@ -209,7 +209,7 @@ test_that("read_feather requires RandomAccessFile and errors nicely otherwise (A
   skip_if_not_available("gzip")
   expect_error(
     read_feather(CompressedInputStream$create(feather_file)),
-    'file must be a "RandomAccessFile"'
+    '`file` must be a RandomAccessFile'
   )
 })
 
@@ -254,7 +254,7 @@ test_that("read_feather closes connection to file", {
   write_feather(tib, sink = tf)
   expect_true(file.exists(tf))
   read_feather(tf)
-  expect_error(file.remove(tf), NA)
+  expect_no_error(file.remove(tf))
   expect_false(file.exists(tf))
 })
 
@@ -333,6 +333,6 @@ test_that("Can read Feather files from a URL", {
   skip_on_cran()
   feather_url <- "https://github.com/apache/arrow-testing/raw/master/data/arrow-ipc-stream/integration/1.0.0-littleendian/generated_datetime.arrow_file" # nolint
   fu <- read_feather(feather_url)
-  expect_true(tibble::is_tibble(fu))
-  expect_identical(dim(fu), c(17L, 15L))
+  expect_s3_class(fu, "tbl_df")
+  expect_shape(fu, dim = c(17L, 15L))
 })
diff --git a/r/tests/testthat/test-field.R b/r/tests/testthat/test-field.R
index e1972180fcea..46f79985d262 100644
--- a/r/tests/testthat/test-field.R
+++ b/r/tests/testthat/test-field.R
@@ -33,7 +33,7 @@ test_that("Field with nullable values", {
 })
 
 test_that("Field validation", {
-  expect_error(schema(b = 32), "b must be a DataType, not numeric")
+  expect_error(schema(b = 32), "`b` must be a DataType")
 })
 
 test_that("Print method for field", {
diff --git a/r/tests/testthat/test-install-arrow.R b/r/tests/testthat/test-install-arrow.R
index 5415fc27581c..2cfadd99bfc1 100644
--- a/r/tests/testthat/test-install-arrow.R
+++ b/r/tests/testthat/test-install-arrow.R
@@ -35,6 +35,5 @@ test_that("arrow_repos", {
 })
 
 test_that("on_rosetta() does not warn", {
-  # There is no warning
-  expect_warning(on_rosetta(), NA)
+  expect_no_warning(on_rosetta())
 })
diff --git a/r/tests/testthat/test-memory-pool.R b/r/tests/testthat/test-memory-pool.R
index e3adeb675fe3..2b71a606b50c 100644
--- a/r/tests/testthat/test-memory-pool.R
+++ b/r/tests/testthat/test-memory-pool.R
@@ -20,7 +20,6 @@ test_that("default_memory_pool and its attributes", {
   # Not integer bc can be >2gb, so we cast to double
   expect_type(pool$bytes_allocated, "double")
   expect_type(pool$max_memory, "double")
-  expect_true(pool$backend_name %in% c("system", "jemalloc", "mimalloc"))
-
-  expect_all_true(supported_memory_backends() %in% c("system", "jemalloc", "mimalloc"))
+  expect_in(pool$backend_name, c("system", "jemalloc", "mimalloc"))
+  expect_in(supported_memory_backends(), c("system", "jemalloc", "mimalloc"))
 })
diff --git a/r/tests/testthat/test-metadata.R b/r/tests/testthat/test-metadata.R
index fea45786357e..f7a30ac38c4f 100644
--- a/r/tests/testthat/test-metadata.R
+++ b/r/tests/testthat/test-metadata.R
@@ -141,13 +141,12 @@ arbitrary\040code\040was\040just\040executed
 254
 "
   )
-  expect_message(
+  expect_no_message(
     expect_warning(
       as.data.frame(tab),
       'Invalid metadata$[["r"]]',
       fixed = TRUE
-    ),
-    NA
+    )
   )
 })
 
@@ -382,9 +381,8 @@ test_that("Row-level metadata (does not) roundtrip in datasets", {
   )
 
   # however there is *no* warning if we don't select the metadata column
-  expect_warning(
-    df_from_ds <- ds |> dplyr::select(int) |> dplyr::collect(),
-    NA
+  expect_no_warning(
+    df_from_ds <- ds |> dplyr::select(int) |> dplyr::collect()
   )
 })
 
diff --git a/r/tests/testthat/test-parquet.R b/r/tests/testthat/test-parquet.R
index faa8d41e23db..7fb7b542ace2 100644
--- a/r/tests/testthat/test-parquet.R
+++ b/r/tests/testthat/test-parquet.R
@@ -22,8 +22,8 @@ pq_file <- system.file("v0.7.1.parquet", package = "arrow")
 test_that("reading a known Parquet file to tibble", {
   skip_if_not_available("snappy")
   df <- read_parquet(pq_file)
-  expect_true(tibble::is_tibble(df))
-  expect_identical(dim(df), c(10L, 11L))
+  expect_s3_class(df, "tbl_df")
+  expect_shape(df, dim = c(10L, 11L))
   # TODO: assert more about the contents
 })
 
@@ -35,7 +35,7 @@ test_that("simple int column roundtrip", {
   df_read <- read_parquet(pq_tmp_file, mmap = FALSE)
   expect_equal(df, df_read)
   # Make sure file connection is cleaned up
-  expect_error(file.remove(pq_tmp_file), NA)
+  expect_no_error(file.remove(pq_tmp_file))
   expect_false(file.exists(pq_tmp_file))
 })
 
@@ -52,7 +52,7 @@ test_that("read_parquet() with raw data", {
   skip_if_not_available("snappy")
   test_raw <- readBin(pq_file, what = "raw", n = 5000)
   df <- read_parquet(test_raw)
-  expect_identical(dim(df), c(10L, 11L))
+  expect_shape(df, dim = c(10L, 11L))
 })
 
 test_that("write_parquet() handles various compression= specs", {
@@ -85,7 +85,7 @@ test_that("write_parquet() handles various use_dictionary= specs", {
   )
   expect_error(
     write_parquet(tab, tempfile(), use_dictionary = 12),
-    "is.logical(use_dictionary) is not TRUE",
+    "`use_dictionary` must be a logical vector",
     fixed = TRUE
   )
 })
@@ -312,7 +312,7 @@ test_that("ParquetFileReader raises an error for non-RandomAccessFile source", {
   skip_if_not_available("gzip")
   expect_error(
     ParquetFileReader$create(CompressedInputStream$create(pq_file)),
-    'file must be a "RandomAccessFile"'
+    '`file` must be a RandomAccessFile'
   )
 })
 
@@ -489,8 +489,8 @@ test_that("Can read Parquet files from a URL", {
   skip_if_not_available("snappy")
   parquet_url <- "https://github.com/apache/arrow/raw/64f2cc7986ce672dd1a8cb268d193617a80a1653/r/inst/v0.7.1.parquet" # nolint
   pu <- read_parquet(parquet_url)
-  expect_true(tibble::is_tibble(pu))
-  expect_identical(dim(pu), c(10L, 11L))
+  expect_s3_class(pu, "tbl_df")
+  expect_shape(pu, dim = c(10L, 11L))
 })
 
 test_that("thrift string and container size can be specified when reading Parquet files", {
diff --git a/r/tests/testthat/test-python-flight.R b/r/tests/testthat/test-python-flight.R
index 2e60957d2ff7..97c665f75c4e 100644
--- a/r/tests/testthat/test-python-flight.R
+++ b/r/tests/testthat/test-python-flight.R
@@ -34,7 +34,7 @@ if (process_is_running("demo_flight_server")) {
     expect_true(flight_obj %in% list_flights(client))
     expect_error(
       flight_put(client, Array$create(c(1:3)), path = flight_obj),
-      regexp = 'data must be a "data.frame", "Table", or "RecordBatch"'
+      regexp = '`data` must be a data frame, Table, or RecordBatch'
     )
   })
 
@@ -48,7 +48,7 @@ if (process_is_running("demo_flight_server")) {
     )
     expect_error(
       flight_put(client, Array$create(c(1:3)), path = flight_obj),
-      regexp = 'data must be a "data.frame", "Table", or "RecordBatch"'
+      regexp = '`data` must be a data frame, Table, or RecordBatch'
     )
   })
 
diff --git a/r/tests/testthat/test-python.R b/r/tests/testthat/test-python.R
index 0da8539ddef4..9626ddf21b84 100644
--- a/r/tests/testthat/test-python.R
+++ b/r/tests/testthat/test-python.R
@@ -30,7 +30,7 @@ test_that("install_pyarrow", {
   venv <- try(reticulate::virtualenv_create("arrow-test"))
   # Bail out if virtualenv isn't available
   skip_if(inherits(venv, "try-error"))
-  expect_error(install_pyarrow("arrow-test", nightly = TRUE), NA)
+  expect_no_error(install_pyarrow("arrow-test", nightly = TRUE))
   # Set this up for the following tests
   reticulate::use_virtualenv("arrow-test")
 })
diff --git a/r/tests/testthat/test-read-record-batch.R b/r/tests/testthat/test-read-record-batch.R
index 7f310e8fc91c..ff004201c390 100644
--- a/r/tests/testthat/test-read-record-batch.R
+++ b/r/tests/testthat/test-read-record-batch.R
@@ -39,7 +39,7 @@ test_that("RecordBatchFileWriter / RecordBatchFileReader roundtrips", {
 
   expect_equal(read_feather(tf, as_data_frame = FALSE, mmap = FALSE), tab)
   # Make sure connections are closed
-  expect_error(file.remove(tf), NA)
+  expect_no_error(file.remove(tf))
   skip_on_os("windows") # This should pass, we've closed the stream
   expect_false(file.exists(tf))
 })
diff --git a/r/tests/testthat/test-s3-minio.R b/r/tests/testthat/test-s3-minio.R
index de1cad547e19..3417cfc682d5 100644
--- a/r/tests/testthat/test-s3-minio.R
+++ b/r/tests/testthat/test-s3-minio.R
@@ -124,7 +124,7 @@ test_that("Confirm s3_bucket works with endpoint_override", {
   os <- bucket$OpenOutputStream("bucket-test.csv")
   write_csv_arrow(example_data, os)
   os$close()
-  expect_true("bucket-test.csv" %in% bucket$ls())
+  expect_contains(bucket$ls(), "bucket-test.csv")
   bucket$DeleteFile("bucket-test.csv")
 })
 
diff --git a/r/tests/testthat/test-s3.R b/r/tests/testthat/test-s3.R
index 7818f1e3d43a..ad8c9727008a 100644
--- a/r/tests/testthat/test-s3.R
+++ b/r/tests/testthat/test-s3.R
@@ -59,7 +59,7 @@ if (run_these) {
   test_that("RandomAccessFile$ReadMetadata() works for S3FileSystem", {
     file <- bucket$OpenInputFile(paste0(now, "/", "test.parquet"))
     metadata <- file$ReadMetadata()
-    expect_true(is.list(metadata))
-    expect_true("ETag" %in% names(metadata))
+    expect_type(metadata, "list")
+    expect_contains(names(metadata), "ETag")
   })
 }
diff --git a/r/tests/testthat/test-schema.R b/r/tests/testthat/test-schema.R
index 93fb16c77fbd..3a18edd98158 100644
--- a/r/tests/testthat/test-schema.R
+++ b/r/tests/testthat/test-schema.R
@@ -131,12 +131,12 @@ test_that("Schema modification", {
   expect_equal(schm, schema(b = double(), y = int64(), d = int8()))
 
   # Error handling
-  expect_error(schm$c <- 4, "value must be a DataType")
-  expect_error(schm[[-3]] <- int32(), "i not greater than 0")
-  expect_error(schm[[0]] <- int32(), "i not greater than 0")
-  expect_error(schm[[NA_integer_]] <- int32(), "!is.na(i) is not TRUE", fixed = TRUE)
-  expect_error(schm[[TRUE]] <- int32(), "i is not a numeric or integer vector")
-  expect_error(schm[[c(2, 4)]] <- int32(), "length(i) not equal to 1", fixed = TRUE)
+  expect_error(schm$c <- 4, "`value` must be a DataType")
+  expect_error(schm[[-3]] <- int32(), "`i` must be")
+  expect_error(schm[[0]] <- int32(), "`i` must be")
+  expect_error(schm[[NA_integer_]] <- int32(), "`i` must be", fixed = TRUE)
+  expect_error(schm[[TRUE]] <- int32(), "`i` must be")
+  expect_error(schm[[c(2, 4)]] <- int32(), "`i` must have length 1", fixed = TRUE)
 })
 
 test_that("Metadata can be reassigned as a whole", {
@@ -326,6 +326,6 @@ test_that("schema print truncation", {
 
   expect_error(
     print_schema_fields(schema(tbl), truncate = TRUE, max_fields = 0),
-    regexp = "max_fields not greater than 0"
+    regexp = "must be a whole number larger than or equal to 1"
   )
 })
diff --git a/r/tests/testthat/test-udf.R b/r/tests/testthat/test-udf.R
index 2eadd87444b5..156edd0b6de9 100644
--- a/r/tests/testthat/test-udf.R
+++ b/r/tests/testthat/test-udf.R
@@ -17,7 +17,7 @@
 
 test_that("list_compute_functions() works", {
   expect_type(list_compute_functions(), "character")
-  expect_all_false(grepl("^hash_", list_compute_functions()))
+  expect_no_match(list_compute_functions(), "^hash_")
 })
 
 test_that("arrow_scalar_function() works", {
@@ -95,8 +95,8 @@ test_that("register_scalar_function() adds a compute function to the registry",
   )
   on.exit(unregister_binding("times_32"))
 
-  expect_true("times_32" %in% names(asNamespace("arrow")$.cache$functions))
-  expect_true("times_32" %in% list_compute_functions())
+  expect_contains(names(asNamespace("arrow")$.cache$functions), "times_32")
+  expect_contains(list_compute_functions(), "times_32")
 
   expect_equal(
     call_function("times_32", Array$create(1L, int32())),