diff --git a/.Rbuildignore b/.Rbuildignore
index e45fa7c85..8bffc965b 100644
--- a/.Rbuildignore
+++ b/.Rbuildignore
@@ -5,10 +5,8 @@
 ^API$
 ^README\.Rmd$
 ^README-.*\.png$
-^.travis.yml
 ^appveyor\.yml$
 ^tic\.R$
-^\.travis\.yml$
 ^docs$
 ^_pkgdown\.yml$
 CONTRIBUTING.md
@@ -20,5 +18,5 @@ revdep
 ^\.pre-commit-config\.yaml$
 ^brew\-log$
 ^\.github/$
-^touchstone/$
-^.github/$
+^touchstone$
+^\.github$
diff --git a/API b/API
index f95f8c1bf..80b09fbe1 100644
--- a/API
+++ b/API
@@ -6,10 +6,11 @@ cache_activate(cache_name = NULL, verbose = TRUE)
 cache_clear(cache_name = NULL, ask = TRUE)
 cache_deactivate(verbose = TRUE)
 cache_info(cache_name = NULL, format = "both")
-create_style_guide(initialize = default_style_guide_attributes, line_break = NULL, space = NULL, token = NULL, indention = NULL, use_raw_indention = FALSE, reindention = tidyverse_reindention(), style_guide_name = NULL, style_guide_version = NULL, more_specs_style_guide = NULL)
+create_style_guide(initialize = default_style_guide_attributes, line_break = NULL, space = NULL, token = NULL, indention = NULL, use_raw_indention = FALSE, reindention = tidyverse_reindention(), style_guide_name = NULL, style_guide_version = NULL, more_specs_style_guide = NULL, transformers_drop = specify_transformers_drop())
 default_style_guide_attributes(pd_flat)
 specify_math_token_spacing(zero = "'^'", one = c("'+'", "'-'", "'*'", "'/'"))
 specify_reindention(regex_pattern = NULL, indention = 0, comments_only = TRUE)
+specify_transformers_drop(spaces = NULL, indention = NULL, line_breaks = NULL, tokens = NULL)
 style_dir(path = ".", ..., style = tidyverse_style, transformers = style(...), filetype = c("R", "Rprofile"), recursive = TRUE, exclude_files = NULL, exclude_dirs = c("packrat", "renv"), include_roxygen_examples = TRUE, base_indention = 0, dry = "off")
 style_file(path, ..., style = tidyverse_style, transformers = style(...), include_roxygen_examples = TRUE, base_indention = 0, dry = "off")
 style_pkg(pkg = ".", ..., style = tidyverse_style, transformers = style(...), filetype = c("R", "Rprofile"), exclude_files = "R/RcppExports.R", exclude_dirs = c("packrat", "renv"), include_roxygen_examples = TRUE, base_indention = 0, dry = "off")
diff --git a/NAMESPACE b/NAMESPACE
index 67f3fcd43..7ab16f529 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -9,6 +9,7 @@ export(create_style_guide)
 export(default_style_guide_attributes)
 export(specify_math_token_spacing)
 export(specify_reindention)
+export(specify_transformers_drop)
 export(style_dir)
 export(style_file)
 export(style_pkg)
diff --git a/NEWS.md b/NEWS.md
index 08eb85409..125839a46 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -21,6 +21,11 @@
   style tokens, you had to always also style spaces, indention, line breaks as 
   well (#705, #707).
 
+- New argument `transformers_drop` in `create_style_guide()` to be populated with
+  new helper function `specify_transformers_drop()` for specifying conditions
+  under which transformers are not going to be used and can therefore be 
+  omitted without effecting the result of styling (#711).
+
 ## Major changes
 
 
@@ -31,9 +36,10 @@
   or focus loss, so this feature becomes less relevant. (#631).
 - blank lines in function calls and headers are now removed, for the former only 
   when there are no comments before or after the blank line (#629, #630, #635).
-- speed improvements: (~10%) when cache is activated because transformers are not 
-  captured as character anymore (#679), ~ 3% in low-level optimization (#691). 
-  Require magrittr 2.0 gives about 7% speed improvement (#681).
+- speed improvements: ~10% when cache is activated because transformers are not 
+  captured as character anymore (#679), ~3% in low-level optimization (#691). 
+  7% by requiring magrittr 2.0 (#681), ~8% by dropping unused transformers 
+  (#711) .
 - `#<<` is now recognized as the xaringan marker and no space is added after`#` 
   (#700).
 
diff --git a/R/rules-spaces.R b/R/rules-spaces.R
index 7c6b7ed12..72f53da72 100644
--- a/R/rules-spaces.R
+++ b/R/rules-spaces.R
@@ -185,20 +185,6 @@ set_space_in_curly_curly <- function(pd) {
   pd
 }
 
-add_space_before_brace <- function(pd_flat) {
-  # TODO remove this, it has no effect since { can only appear in the first
-  # position of the nest and taking lead(op_after, default = FALSE) will always
-  # yield a vector of FALSE only.
-  op_after <- pd_flat$token %in% "'{'"
-  if (!any(op_after)) {
-    return(pd_flat)
-  }
-  op_before <- lead(op_after, default = FALSE)
-  idx_before <- op_before & (pd_flat$newlines == 0L) & pd_flat$token != "'('"
-  pd_flat$spaces[idx_before] <- pmax(pd_flat$spaces[idx_before], 1L)
-  pd_flat
-}
-
 add_space_after_comma <- function(pd_flat) {
   comma_after <- (pd_flat$token == "','") & (pd_flat$newlines == 0L)
   pd_flat$spaces[comma_after] <- pmax(pd_flat$spaces[comma_after], 1L)
diff --git a/R/style-guides.R b/R/style-guides.R
index aeb3140d7..5fae0f40c 100644
--- a/R/style-guides.R
+++ b/R/style-guides.R
@@ -74,25 +74,23 @@ tidyverse_style <- function(scope = "tokens",
 
 
   indention_manipulators <- if ("indention" %in% scope) {
-    lst(
+    list(
       indent_braces = partial(indent_braces, indent_by = indent_by),
-      unindent_fun_dec,
+      unindent_fun_dec = unindent_fun_dec,
       indent_op = partial(indent_op, indent_by = indent_by),
       indent_eq_sub = partial(indent_eq_sub, indent_by = indent_by),
       indent_without_paren = partial(indent_without_paren,
         indent_by = indent_by
       ),
-      update_indention_ref_fun_dec =
-        if ("indention" %in% scope) update_indention_ref_fun_dec
+      update_indention_ref_fun_dec = update_indention_ref_fun_dec
     )
   }
   space_manipulators <- if ("spaces" %in% scope) {
-    lst(
-      remove_space_before_closing_paren,
+    list(
+      remove_space_before_closing_paren = remove_space_before_closing_paren,
       remove_space_before_opening_paren = if (strict) remove_space_before_opening_paren,
-      add_space_after_for_if_while,
-      add_space_before_brace,
-      remove_space_before_comma,
+      add_space_after_for_if_while = add_space_after_for_if_while,
+      remove_space_before_comma = remove_space_before_comma,
       style_space_around_math_token = partial(
         style_space_around_math_token, strict,
         math_token_spacing$zero,
@@ -105,34 +103,34 @@ tidyverse_style <- function(scope = "tokens",
       spacing_around_op = purrr::partial(set_space_around_op,
         strict = strict
       ),
-      remove_space_after_opening_paren,
-      remove_space_after_excl,
-      set_space_after_bang_bang,
-      remove_space_before_dollar,
-      remove_space_after_fun_dec,
-      remove_space_around_colons,
+      remove_space_after_opening_paren = remove_space_after_opening_paren,
+      remove_space_after_excl = remove_space_after_excl,
+      set_space_after_bang_bang = set_space_after_bang_bang,
+      remove_space_before_dollar = remove_space_before_dollar,
+      remove_space_after_fun_dec = remove_space_after_fun_dec,
+      remove_space_around_colons = remove_space_around_colons,
       start_comments_with_space = partial(start_comments_with_space,
         force_one = start_comments_with_one_space
       ),
-      remove_space_after_unary_pm_nested,
+      remove_space_after_unary_pm_nested = remove_space_after_unary_pm_nested,
       spacing_before_comments = if (strict) {
         set_space_before_comments
       } else {
         add_space_before_comments
       },
-      set_space_between_levels,
-      set_space_between_eq_sub_and_comma,
-      set_space_in_curly_curly
+      set_space_between_levels = set_space_between_levels,
+      set_space_between_eq_sub_and_comma = set_space_between_eq_sub_and_comma,
+      set_space_in_curly_curly = set_space_in_curly_curly
     )
   }
 
   use_raw_indention <- !("indention" %in% scope)
 
   line_break_manipulators <- if ("line_breaks" %in% scope) {
-    lst(
-      set_line_break_around_comma_and_or,
-      set_line_break_after_assignment,
-      set_line_break_before_curly_opening,
+    list(
+      set_line_break_around_comma_and_or = set_line_break_around_comma_and_or,
+      set_line_break_after_assignment = set_line_break_after_assignment,
+      set_line_break_before_curly_opening = set_line_break_before_curly_opening,
       remove_line_break_before_round_closing_after_curly =
         if (strict) remove_line_break_before_round_closing_after_curly,
       remove_line_breaks_in_fun_dec =
@@ -143,7 +141,7 @@ tidyverse_style <- function(scope = "tokens",
       ),
       # must be after style_line_break_around_curly as it remove line
       # breaks again for {{.
-      set_line_break_around_curly_curly,
+      set_line_break_around_curly_curly = set_line_break_around_curly_curly,
       set_line_break_after_opening_if_call_is_multi_line = if (strict) {
         partial(
           set_line_break_after_opening_if_call_is_multi_line,
@@ -157,24 +155,77 @@ tidyverse_style <- function(scope = "tokens",
           except_token_before = "COMMENT"
         )
       },
-      purrr::partial(remove_line_break_in_fun_call, strict = strict),
+      remove_line_break_in_fun_call = purrr::partial(remove_line_break_in_fun_call, strict = strict),
       add_line_break_after_pipe = if (strict) add_line_break_after_pipe,
       set_linebreak_after_ggplot2_plus = if (strict) set_linebreak_after_ggplot2_plus
     )
   }
 
   token_manipulators <- if ("tokens" %in% scope) {
-    lst(
-      fix_quotes,
-      force_assignment_op,
-      resolve_semicolon,
-      add_brackets_in_pipe,
-      remove_terminal_token_before_and_after,
+    list(
+      fix_quotes = fix_quotes,
+      force_assignment_op = force_assignment_op,
+      resolve_semicolon = resolve_semicolon,
+      add_brackets_in_pipe = add_brackets_in_pipe,
+      remove_terminal_token_before_and_after = remove_terminal_token_before_and_after,
       wrap_if_else_while_for_fun_multi_line_in_curly =
         if (strict) wrap_if_else_while_for_fun_multi_line_in_curly
     )
   }
 
+  transformers_drop <- specify_transformers_drop(
+    spaces = list(
+      # remove_space_before_closing_paren = c("')'", "']'"),
+      # remove_space_before_opening_paren = c("'('", "'['", "LBB"),
+      add_space_after_for_if_while = c("IF", "WHILE", "FOR"),
+      # remove_space_before_comma = "','",
+      set_space_between_eq_sub_and_comma = "EQ_SUB",
+      style_space_around_math_token = c(
+        math_token_spacing$zero,
+        math_token_spacing$one
+      ),
+      style_space_around_tilde = "'~'",
+      # remove_space_after_opening_paren = c("'('", "'['", "LBB"),
+      remove_space_after_excl = "'!'",
+      set_space_after_bang_bang = "'!'",
+      remove_space_before_dollar = "'$'",
+      remove_space_after_fun_dec = "FUNCTION",
+      remove_space_around_colons = c("':'", "NS_GET_INT", "NS_GET"),
+      start_comments_with_space = "COMMENT",
+      remove_space_after_unary_pm_nested = c("'+'", "'-'"),
+      spacing_before_comments = "COMMENT",
+      set_space_in_curly_curly = c("'{'", "'}'")
+    ),
+    indention = list(
+      # indent_braces = c("'('", "'['", "'{'", "')'", "']'", "'}'"),
+      unindent_fun_dec = "FUNCTION",
+      indent_eq_sub = c("EQ_SUB", "EQ_FORMALS"), # TODO rename
+      update_indention_ref_fun_dec = "FUNCTION"
+    ),
+    line_breaks = list(
+      set_line_break_before_curly_opening = "'{'",
+      remove_line_break_before_round_closing_after_curly = "'}'",
+      remove_line_breaks_in_fun_dec = "FUNCTION",
+      set_line_break_around_curly_curly = "'{'",
+      style_line_break_around_curly = "'{'",
+      add_line_break_after_pipe = "SPECIAL-PIPE"
+    ),
+    tokens = list(
+      resolve_semicolon = "';'",
+      add_brackets_in_pipe = "SPECIAL-PIPE",
+      # before 3.6, these assignments are not wrapped into top level expression
+      # and `text` supplied to transformers_drop() is "", so it appears to not
+      # contain EQ_ASSIGN, and the transformer is falsely removed.
+      # compute_parse_data_nested / text_to_flat_pd ('a = 4')
+      force_assignment_op = "EQ_ASSIGN",
+      wrap_if_else_while_for_fun_multi_line_in_curly = c("IF", "WHILE", "FOR", "FUNCTION")
+    )
+  )
+
+  if (getRversion() < 3.6) {
+    transformers_drop$token$force_assignment_op <- NULL
+  }
+
   style_guide_name <- "styler::tidyverse_style@https://github.com/r-lib"
   create_style_guide(
     # transformer functions
@@ -188,7 +239,8 @@ tidyverse_style <- function(scope = "tokens",
     reindention            =                    reindention,
     style_guide_name       =               style_guide_name,
     style_guide_version    =                 styler_version,
-    more_specs_style_guide =                           args
+    more_specs_style_guide =                           args,
+    transformers_drop      =              transformers_drop
   )
 }
 
@@ -231,6 +283,14 @@ tidyverse_style <- function(scope = "tokens",
 #'   they will yield generic code and we loose the specific value of `arg` (see
 #'   [styler::cache_make_key()]), even when unquoting these inputs with `!!`
 #'   beforehand in `purrr::partial()`.
+#' @param transformers_drop A list specifying under which conditions
+#'   transformer functions can be dropped since they have no effect on the
+#'   code to format, most easily constructed with
+#'   [specify_transformers_drop()]. This is argument experimental and may
+#'   change in future releases without prior notification. It was mainly
+#'   introduced to improve speed. Listing transformers here that occur almost
+#'   always in code does not make sense because the process of excluding them
+#'   also takes some time.
 #' @examples
 #' set_line_break_before_curly_opening <- function(pd_flat) {
 #'   op <- pd_flat$token %in% "'{'"
@@ -257,24 +317,89 @@ create_style_guide <- function(initialize = default_style_guide_attributes,
                                reindention = tidyverse_reindention(),
                                style_guide_name = NULL,
                                style_guide_version = NULL,
-                               more_specs_style_guide = NULL) {
-  lst(
+                               more_specs_style_guide = NULL,
+                               transformers_drop = specify_transformers_drop()) {
+  list(
     # transformer functions
-    initialize = lst(initialize),
-    line_break,
-    space,
-    token,
-    indention,
+    initialize = list(initialize = initialize),
+    line_break = line_break,
+    space = space,
+    token = token,
+    indention = indention,
     # transformer options
-    use_raw_indention,
-    reindention,
-    style_guide_name,
-    style_guide_version,
-    more_specs_style_guide
+    use_raw_indention = use_raw_indention,
+    reindention = reindention,
+    style_guide_name = style_guide_name,
+    style_guide_version = style_guide_version,
+    more_specs_style_guide = more_specs_style_guide,
+    transformers_drop = transformers_drop
   ) %>%
     map(compact)
 }
 
+#' Specify which tokens must be absent for a transformer to be dropped
+#'
+#' `{styler}` can remove transformer functions safely removed from the list of
+#' transformers to be applied on every *nest* with [transformers_drop()] if the
+#' tokens that trigger a manipulation of the parse data are absent in the text
+#' to style. `specify_transformers_drop()` helps you specify these
+#' conditions.
+#'
+#' Note that the negative formulation (must be absent in order to be dropped)
+#' means that when you add a new rule and you forget
+#' to add a rule for when to drop it, it will not be dropped. If we required to
+#' specify the complement (which tokens must be present for the transformer to
+#' be kept), the transformer would be silently removed, which is less save.
+#' @param spaces,indention,line_breaks,tokens Each a list (or `NULL`) where
+#'   the name of each element is the concerning transformer, the value is an
+#'   unnamed vector with tokens that match the rule. See 'Examples'.
+#'
+#' @section Warning:
+#' It is the responsibility of the developer to ensure expected behavior, in
+#' particular that:
+#' * the name of the supplied dropping criteria matches the name of the
+#'   transformer function.
+#' * the dropping criteria (name + token) reflects correctly under which
+#'   circumstances the transformer does not have an impact on styling and can
+#'   therefore be safely removed without affecting the styling outcome.
+#'
+#' You can use the unexported function [test_transformers_drop()] for some
+#' checks.
+#' @examples
+#' dropping <- specify_transformers_drop(
+#'   spaces = c(remove_space_after_excl = "'!'")
+#' )
+#' style_guide <- create_style_guide(
+#'   space = list(remove_space_after_excl = styler:::remove_space_after_excl),
+#'   transformers_drop = dropping
+#' )
+#' # transformers_drop() will remove the transformer when the code does not
+#' # contain an exclamation mark
+#' style_guide_with_some_transformers_dropped <- styler:::transformers_drop(
+#'   "x <- 3;2", style_guide
+#' )
+#' setdiff(
+#'   names(style_guide$space),
+#'   names(style_guide_with_some_transformers_dropped)
+#' )
+#' # note that dropping all transformers of a scope means that this scope
+#' # has an empty named list for this scope
+#' style_guide_with_some_transformers_dropped$space
+#' # this is not the same as if this scope was never specified.
+#' tidyverse_style(scope = "none")$space
+#' # Hence, styler should check for length 0 to decide if a scope is present or
+#' # not, not via `is.null()` and we can use the `is.null()` check to see if
+#' # this scope was initially required by the user.
+#' @export
+specify_transformers_drop <- function(spaces = NULL,
+                                      indention = NULL,
+                                      line_breaks = NULL,
+                                      tokens = NULL) {
+  list(
+    space = spaces, indention = indention, line_break = line_breaks,
+    token = tokens
+  )
+}
 
 #' Specify what is re-indented how
 #'
@@ -302,10 +427,10 @@ NULL
 specify_reindention <- function(regex_pattern = NULL,
                                 indention = 0,
                                 comments_only = TRUE) {
-  lst(
-    regex_pattern,
-    indention,
-    comments_only
+  list(
+    regex_pattern = regex_pattern,
+    indention = indention,
+    comments_only = comments_only
   )
 }
 
@@ -381,9 +506,9 @@ specify_math_token_spacing <-
   function(zero = "'^'",
            one = c("'+'", "'-'", "'*'", "'/'")) {
     assert_tokens(c(one, zero))
-    lst(
+    list(
       one = setdiff(c(math_token, one), zero),
-      zero
+      zero = zero
     )
   }
 
diff --git a/R/testing.R b/R/testing.R
index 6d7c90883..b5b00f0d1 100644
--- a/R/testing.R
+++ b/R/testing.R
@@ -341,3 +341,28 @@ fresh_testthat_cache <- function() {
 cache_more_specs_default <- function() {
   cache_more_specs(include_roxygen_examples = TRUE, base_indention = 0)
 }
+
+#' Test `transformers_drop` for consistency
+#'
+#' Check if the argument `transformers_drop` in [create_style_guide()] is
+#' consistent with the transformers specified in that function.
+#' @param transformers The output of [create_style_guide()] we want to test.
+#' @keywords internal
+test_transformers_drop <- function(transformers) {
+  scopes <- intersect(
+    names(transformers$transformers_drop),
+    names(transformers)
+  )
+
+  purrr::walk2(transformers$transformers_drop, transformers[scopes], function(x, y) {
+    # all x must be in y. select the x that are not in y
+    diff <- setdiff(names(x), names(y))
+    if (length(diff) > 0) {
+      rlang::abort(paste(
+        "transformers_drop specifies exclusion rules for transformers that ",
+        "are not in the style guilde. Please add the rule to the style guide ",
+        "or remove the dropping rules:", paste(diff, collapse = ", ")
+      ))
+    }
+  })
+}
diff --git a/R/transform-code.R b/R/transform-code.R
index eefcaba9d..eeef0856c 100644
--- a/R/transform-code.R
+++ b/R/transform-code.R
@@ -64,7 +64,7 @@ separate_chunks <- function(lines, filetype) {
     c(1, r_raw_chunks$ends), c(r_raw_chunks$starts, length(lines)),
     ~ lines[seq2(.x, .y)]
   )
-  lst(r_chunks, text_chunks)
+  list(r_chunks = r_chunks, text_chunks = text_chunks)
 }
 
 #' Identifies raw Rmd or Rnw code chunks
diff --git a/R/transform-files.R b/R/transform-files.R
index 7d266b3b9..cccae05ec 100644
--- a/R/transform-files.R
+++ b/R/transform-files.R
@@ -169,6 +169,7 @@ parse_transform_serialize_roxygen <- function(text, transformers, base_indention
     flatten_chr()
 }
 
+
 #' Split text into roxygen and non-roxygen example segments
 #'
 #' @param text Roxygen comments
@@ -186,7 +187,7 @@ parse_transform_serialize_roxygen <- function(text, transformers, base_indention
 #' @keywords internal
 split_roxygen_segments <- function(text, roxygen_examples) {
   if (is.null(roxygen_examples)) {
-    return(lst(separated = list(text), selectors = NULL))
+    return(list(separated = list(text), selectors = NULL))
   }
   all_lines <- seq2(1L, length(text))
   active_segment <- as.integer(all_lines %in% roxygen_examples)
@@ -194,7 +195,7 @@ split_roxygen_segments <- function(text, roxygen_examples) {
   separated <- split(text, factor(segment_id))
   restyle_selector <- ifelse(roxygen_examples[1] == 1L, odd_index, even_index)
 
-  lst(separated, selectors = restyle_selector(separated))
+  list(separated = separated, selectors = restyle_selector(separated))
 }
 
 #' Parse, transform and serialize text
@@ -217,18 +218,21 @@ parse_transform_serialize_r <- function(text,
 
   text <- assert_text(text)
   pd_nested <- compute_parse_data_nested(text, transformers, more_specs)
-  blank_lines_to_next_expr <- find_blank_lines_to_next_block(pd_nested)
   if (nrow(pd_nested) == 0) {
     if (warn_empty) {
       warn("Text to style did not contain any tokens. Returning empty string.")
     }
     return("")
   }
+  transformers <- transformers_drop(
+    if (getRversion() < 3.4) text else pd_nested$text[!pd_nested$is_cached],
+    transformers
+  )
 
   text_out <- pd_nested %>%
     split(pd_nested$block) %>%
     unname() %>%
-    map2(blank_lines_to_next_expr,
+    map2(find_blank_lines_to_next_block(pd_nested),
       parse_transform_serialize_r_block,
       transformers = transformers,
       base_indention = base_indention
@@ -245,6 +249,38 @@ parse_transform_serialize_r <- function(text,
   text_out
 }
 
+
+#' Remove transformers that are not needed
+#'
+#' The goal is to speed up styling by removing all rules that are only
+#' applicable in contexts that don't occur often, e.g. for most code, we don't
+#' expect ";" to be in it, so we don't need to apply `resolve_semicolon()` on
+#' every *nest*.
+#' @param text Text to parse. Can also be the column `text` of the output of
+#'   [compute_parse_data_nested()], where each element is a token (instead of a
+#'   line).
+#' @param transformers the transformers.
+#' @keywords internal
+#' @seealso specify_transformers_drop
+transformers_drop <- function(text, transformers) {
+  is_colon <- text == ";"
+  if (any(is_colon)) {
+    # ; can only be parsed when on the same line as other token, not the case
+    # here since text is output of compute_parse_data_nested.
+    text <- c(text[!is_colon], "1;")
+  }
+  token <- unique(tokenize(text)$token)
+  for (scope in c("line_break", "space", "token", "indention")) {
+    rules <- transformers$transformers_drop[[scope]]
+    for (rule in names(rules)) {
+      if (!any(rules[[rule]] %in% token)) {
+        transformers[[scope]][rule] <- NULL
+      }
+    }
+  }
+  transformers
+}
+
 #' Apply transformers to a parse table
 #'
 #' The column `multi_line` is updated (after the line break information is
@@ -273,7 +309,7 @@ apply_transformers <- function(pd_nested, transformers) {
     pd_nested,
     c(
       transformers$initialize, transformers$line_break, set_multi_line,
-      if (!is.null(transformers$line_break)) update_newlines
+      if (length(transformers$line_break) != 0) update_newlines
     )
   )
 
@@ -303,7 +339,7 @@ apply_transformers <- function(pd_nested, transformers) {
 #'   Needed for reverse engineering the scope.
 #' @keywords internal
 can_verify_roundtrip <- function(transformers) {
-  is.null(transformers$token)
+  length(transformers$token) == 0
 }
 
 #' Verify the styling
diff --git a/inst/WORDLIST b/inst/WORDLIST
index 957c3a8b0..0b4fe44db 100644
--- a/inst/WORDLIST
+++ b/inst/WORDLIST
@@ -69,6 +69,7 @@ href
 http
 https
 icloud
+impl
 infinitively
 initializer
 innode
@@ -193,6 +194,7 @@ tidyeval
 tidyr
 tidyverse
 Tidyverse
+tokenized
 travis
 tryCatch
 tryGugus
diff --git a/man/create_style_guide.Rd b/man/create_style_guide.Rd
index bb07f00e3..ef1e8add1 100644
--- a/man/create_style_guide.Rd
+++ b/man/create_style_guide.Rd
@@ -14,7 +14,8 @@ create_style_guide(
   reindention = tidyverse_reindention(),
   style_guide_name = NULL,
   style_guide_version = NULL,
-  more_specs_style_guide = NULL
+  more_specs_style_guide = NULL,
+  transformers_drop = specify_transformers_drop()
 )
 }
 \arguments{
@@ -55,6 +56,15 @@ such functions are converted to characters in \code{\link[=cache_make_key]{cache
 they will yield generic code and we loose the specific value of \code{arg} (see
 \code{\link[=cache_make_key]{cache_make_key()}}), even when unquoting these inputs with \verb{!!}
 beforehand in \code{purrr::partial()}.}
+
+\item{transformers_drop}{A list specifying under which conditions
+transformer functions can be dropped since they have no effect on the
+code to format, most easily constructed with
+\code{\link[=specify_transformers_drop]{specify_transformers_drop()}}. This is argument experimental and may
+change in future releases without prior notification. It was mainly
+introduced to improve speed. Listing transformers here that occur almost
+always in code does not make sense because the process of excluding them
+also takes some time.}
 }
 \description{
 This is a helper function to create a style guide, which is technically
diff --git a/man/specify_transformers_drop.Rd b/man/specify_transformers_drop.Rd
new file mode 100644
index 000000000..e9069cc33
--- /dev/null
+++ b/man/specify_transformers_drop.Rd
@@ -0,0 +1,74 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/style-guides.R
+\name{specify_transformers_drop}
+\alias{specify_transformers_drop}
+\title{Specify which tokens must be absent for a transformer to be dropped}
+\usage{
+specify_transformers_drop(
+  spaces = NULL,
+  indention = NULL,
+  line_breaks = NULL,
+  tokens = NULL
+)
+}
+\arguments{
+\item{spaces, indention, line_breaks, tokens}{Each a list (or \code{NULL}) where
+the name of each element is the concerning transformer, the value is an
+unnamed vector with tokens that match the rule. See 'Examples'.}
+}
+\description{
+\code{{styler}} can remove transformer functions safely removed from the list of
+transformers to be applied on every \emph{nest} with \code{\link[=transformers_drop]{transformers_drop()}} if the
+tokens that trigger a manipulation of the parse data are absent in the text
+to style. \code{specify_transformers_drop()} helps you specify these
+conditions.
+}
+\details{
+Note that the negative formulation (must be absent in order to be dropped)
+means that when you add a new rule and you forget
+to add a rule for when to drop it, it will not be dropped. If we required to
+specify the complement (which tokens must be present for the transformer to
+be kept), the transformer would be silently removed, which is less save.
+}
+\section{Warning}{
+
+It is the responsibility of the developer to ensure expected behavior, in
+particular that:
+\itemize{
+\item the name of the supplied dropping criteria matches the name of the
+transformer function.
+\item the dropping criteria (name + token) reflects correctly under which
+circumstances the transformer does not have an impact on styling and can
+therefore be safely removed without affecting the styling outcome.
+}
+
+You can use the unexported function \code{\link[=test_transformers_drop]{test_transformers_drop()}} for some
+checks.
+}
+
+\examples{
+dropping <- specify_transformers_drop(
+  spaces = c(remove_space_after_excl = "'!'")
+)
+style_guide <- create_style_guide(
+  space = list(remove_space_after_excl = styler:::remove_space_after_excl),
+  transformers_drop = dropping
+)
+# transformers_drop() will remove the transformer when the code does not
+# contain an exclamation mark
+style_guide_with_some_transformers_dropped <- styler:::transformers_drop(
+  "x <- 3;2", style_guide
+)
+setdiff(
+  names(style_guide$space),
+  names(style_guide_with_some_transformers_dropped)
+)
+# note that dropping all transformers of a scope means that this scope
+# has an empty named list for this scope
+style_guide_with_some_transformers_dropped$space
+# this is not the same as if this scope was never specified.
+tidyverse_style(scope = "none")$space
+# Hence, styler should check for length 0 to decide if a scope is present or
+# not, not via `is.null()` and we can use the `is.null()` check to see if
+# this scope was initially required by the user.
+}
diff --git a/man/test_transformers_drop.Rd b/man/test_transformers_drop.Rd
new file mode 100644
index 000000000..bd2bfa381
--- /dev/null
+++ b/man/test_transformers_drop.Rd
@@ -0,0 +1,16 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/testing.R
+\name{test_transformers_drop}
+\alias{test_transformers_drop}
+\title{Test \code{transformers_drop} for consistency}
+\usage{
+test_transformers_drop(transformers)
+}
+\arguments{
+\item{transformers}{The output of \code{\link[=create_style_guide]{create_style_guide()}} we want to test.}
+}
+\description{
+Check if the argument \code{transformers_drop} in \code{\link[=create_style_guide]{create_style_guide()}} is
+consistent with the transformers specified in that function.
+}
+\keyword{internal}
diff --git a/man/transformers_drop.Rd b/man/transformers_drop.Rd
new file mode 100644
index 000000000..19623db73
--- /dev/null
+++ b/man/transformers_drop.Rd
@@ -0,0 +1,25 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/transform-files.R
+\name{transformers_drop}
+\alias{transformers_drop}
+\title{Remove transformers that are not needed}
+\usage{
+transformers_drop(text, transformers)
+}
+\arguments{
+\item{text}{Text to parse. Can also be the column \code{text} of the output of
+\code{\link[=compute_parse_data_nested]{compute_parse_data_nested()}}, where each element is a token (instead of a
+line).}
+
+\item{transformers}{the transformers.}
+}
+\description{
+The goal is to speed up styling by removing all rules that are only
+applicable in contexts that don't occur often, e.g. for most code, we don't
+expect ";" to be in it, so we don't need to apply \code{resolve_semicolon()} on
+every \emph{nest}.
+}
+\seealso{
+specify_transformers_drop
+}
+\keyword{internal}
diff --git a/tests/testthat/test-serialize_tests.R b/tests/testthat/test-serialize_tests.R
index 19faf3f9b..77433ee1d 100644
--- a/tests/testthat/test-serialize_tests.R
+++ b/tests/testthat/test-serialize_tests.R
@@ -26,3 +26,26 @@ test_that("properly detects match", {
     "identical"
   )
 })
+
+test_that('detects non-matching style guides', {
+  sg <- create_style_guide(
+    space = list(
+      a1 = function(...) NULL,
+      b1 = function(... ) 1
+    ),
+    transformers_drop = specify_transformers_drop(
+      spaces = c(a1 = "'+'")
+    )
+  )
+  expect_silent(test_transformers_drop(sg))
+
+  sg <- create_style_guide(
+    space = list(
+      a1 = function(...) NULL
+    ),
+    transformers_drop = specify_transformers_drop(
+      spaces = c(a2 = "'+'")
+    )
+  )
+  expect_error(test_transformers_drop(sg))
+})
diff --git a/tests/testthat/test-transformers-drop.R b/tests/testthat/test-transformers-drop.R
new file mode 100644
index 000000000..a93aa4921
--- /dev/null
+++ b/tests/testthat/test-transformers-drop.R
@@ -0,0 +1,147 @@
+# c/cp from remove_space_after_excl but for self-containement repeated
+remove_space_after_excl_ <- function(pd_flat) {
+  excl <- (pd_flat$token == "'!'") &
+    (pd_flat$token_after != "'!'") &
+    (pd_flat$newlines == 0L)
+  pd_flat$spaces[excl] <- 0L
+  pd_flat
+}
+
+t <- create_style_guide(
+  space = lst(remove_space_after_excl_),
+  transformers_drop = specify_transformers_drop(
+    spaces = list(remove_space_after_excl_ = c("'!'"))
+  ),
+  style_guide_name = "styler::t@https://github.com/r-lib",
+  style_guide_version = as.character(packageVersion("styler"))
+)
+
+t_no_drop <- create_style_guide(
+  space = lst(remove_space_after_excl_),
+  transformers_drop = NULL,
+)
+
+t_empty_drop1 <- create_style_guide(
+  space = lst(remove_space_after_excl_),
+  transformers_drop = list(space = list()),
+)
+
+t_empty_drop2 <- create_style_guide(
+  space = lst(remove_space_after_excl_),
+  transformers_drop = list(),
+)
+
+test_that("transformers are not removed if they are used", {
+  t_new <- transformers_drop(
+    "!x", t
+  )
+  expect_equal(t_new, t)
+})
+
+test_that("transformers are removed if they are unused", {
+  t_fun <- transformers_drop(
+    "x", t
+  )
+  t_manual <- t
+  t_manual$space$remove_space_after_excl_ <- NULL
+  expect_equal(t_fun, t_manual)
+})
+
+test_that("tidyverse transformers are correctly named", {
+  # test that all dropping rules match an actual rule in the style guide
+  expect_silent(
+    test_transformers_drop(tidyverse_style())
+  )
+})
+
+test_that("tidyverse transformers are correctly dropped", {
+  # TODO maybe there is a more minimal test than this.
+  t_style <- tidyverse_style()
+  t_fun <- transformers_drop("x", t_style)
+
+  names_line_break <- c(
+    "set_line_break_around_comma_and_or",
+    "set_line_break_after_assignment",
+    "set_line_break_after_opening_if_call_is_multi_line",
+    "set_line_break_before_closing_call",
+    "remove_line_break_in_fun_call",
+    "set_linebreak_after_ggplot2_plus"
+  )
+  expect_setequal(names(t_fun$line_break), names_line_break)
+
+  names_spaces <- c(
+    "remove_space_before_closing_paren",
+    "remove_space_before_opening_paren",
+    "remove_space_before_comma",
+    "spacing_around_op",
+    "remove_space_after_opening_paren",
+    "set_space_between_levels"
+  )
+
+  expect_setequal(names(t_fun$space), names_spaces)
+
+  names_indention <- c("indent_braces", "indent_op", "indent_without_paren")
+  expect_setequal(names(t_fun$indention), names_indention)
+
+  names_tokens <- c(
+    "fix_quotes",
+    if (getRversion() < 3.6) "force_assignment_op",
+    "remove_terminal_token_before_and_after"
+  )
+  expect_setequal(names(t_fun$token), names_tokens)
+})
+
+
+test_that("if no transformers_drop is specified, no transformer is removed and no error issued", {
+  t_fun <- transformers_drop(
+    "x", t_no_drop
+  )
+  expect_equal(t_fun, t_no_drop)
+
+  t_fun <- transformers_drop(
+    "x", t_empty_drop1
+  )
+  expect_equal(t_fun, t_empty_drop1)
+
+  t_fun <- transformers_drop(
+    "x", t_empty_drop2
+  )
+  expect_equal(t_fun, t_empty_drop2)
+})
+
+test_that('semi-colon is parsed without error', {
+  expect_equal(
+    transformers_drop(c("!a", ";", "b"), t),
+    t
+  )
+})
+
+
+test_that('can handle old style guide without transformer object', {
+  t_new <- t
+  t_new$transformers_drop <- NULL
+  expect_error(
+    transformers_drop(c("!a", ";", "b"), t_new),
+    NA
+  )
+  expect_error(
+    style_text('1;3', transformers = t_new),
+    NA
+  )
+})
+
+test_that("can handle default", {
+  t_no_drop <- create_style_guide(
+    space = lst(remove_space_after_excl_),
+    style_guide_name = "styler::t@https://github.com/r-lib",
+    style_guide_version = as.character(packageVersion("styler"))
+  )
+  expect_error(
+    transformers_drop(c("!a", ";", "b"), t_no_drop),
+    NA
+  )
+  expect_error(
+    style_text('a =2 ', transformers = t_no_drop),
+    NA
+  )
+})